One read of Claude Developer Platformapi-20260930T153709Z
266 pages moved out of 642 read.
Pages moved
266
significant first
Pages read
642
in this capture
Captured
15:37 UTC
Corpus hash
5fbd08954ac7
corpus-hash
What this read moved
1-25 of 266, page 1 of 11This capture is too large to show at once. Changes 1-25 of 266 are below, significant first; the rest are on the following screens.
agents-and-tools/agent-skills/claude-api-skill Changed · +22 / -1 lines
## Checking an integration for preserved thinking
from line 27
2727* **Batch processing:** Offline batch processing at 50% cost
2828* **Prompt caching:** Prefix-stability design, breakpoint placement, and silent-invalidator audit
2929* **Model migration:** Step-by-step guidance for migrating to newer Claude models (including the breaking changes and behavior shifts on [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-5), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/migration-guide#migrating-from-claude-sonnet-5), and [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/migration-guide))
30* **Preserved thinking migration:** Finding, measuring, and fixing the history edits that invalidate `thinking` blocks, available through the `/claude-api preserved-thinking-migration` subcommand
3031* **Current model information:** Model IDs, context window sizes, and pricing
3132* **Common pitfalls:** Detailed guidance on avoiding frequent mistakes when integrating with the API
3233
from line 121
120121* **Cloud platform detection**, preserving platform-specific model ID formats (for example, the `anthropic.` prefix on Amazon Bedrock) and skipping changes for features that are unavailable on partner-operated platforms
121122* **Breaking parameter changes**, such as removing `temperature`, `top_p`, and `top_k` for Claude Opus 4.8 and Claude Opus 4.7, and converting `thinking: {type: "enabled", budget_tokens: N}` to `thinking: {type: "adaptive"}`
122123* **Prefill replacement**, converting assistant-message prefill patterns to [structured outputs](https://platform.claude.com/docs/en/build-with-claude/structured-outputs) where applicable
123* **Beta header cleanup**, removing beta headers that the target model doesn't require (for example, `effort-2025-11-24`, `fine-grained-tool-streaming-2025-05-14`, `interleaved-thinking-2025-05-14`) and switching back from `client.beta.messages.create` to `client.messages.create`
124* **Beta header cleanup**, removing beta headers that the target model doesn't require (for example, `effort-2025-11-24`, `fine-grained-tool-streaming-2025-05-14`, `interleaved-thinking-2025-05-14`) and switching back from `client.beta.messages.create()` (python, typescript, ruby; go: `client.Beta.Messages.New()`; java: `client.beta().messages().create()`; csharp: `client.Beta.Messages.Create()`; php: `$client->beta->messages->create()`) to `client.messages.create()` (python, typescript, ruby; go: `client.Messages.New()`; java: `client.messages().create()`; csharp: `client.Messages.Create()`; php: `$client->messages->create()`)
124125* **Effort calibration**, recommending an `output_config.effort` starting point for the target model (for example, the default `high` on Claude Opus 5, and `xhigh` for coding and agentic use cases on Claude Opus 4.8 and Claude Opus 4.7)
125126* **Prompt-behavior tuning**, flagging length-control, tool-triggering, subagent, and instruction-following prompts that may behave differently on the target model
126127* **Silent default handling**, opting back into thinking summarization (`thinking.display: "summarized"`) when reasoning is surfaced to users on Claude Opus 4.8 and Claude Opus 4.7
from line 130
129130As it edits, the skill explains each change and its motivation inline. On completion, it produces a checklist of items that require manual verification (typically integration tests, length-control prompt tuning, and cost/rate-limit re-baselining).
130131
131132For the full list of model-specific changes the skill applies, see [Migrating to Claude Opus 5.5 from Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-5), [Migrating to Claude Opus 5.5 from Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-4-8), [Migrating to Claude Sonnet 5.5 from Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5-5/migration-guide#migrating-from-claude-sonnet-5), and [Migrating to Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/migration-guide).
133
134## Checking an integration for preserved thinking
135
136[Preserved thinking](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking) keeps a `thinking` block valid only while the `system` prompt, `tools`, and messages before it stay unchanged. To find and fix the places where your code edits that history, open the repository that builds your requests and invoke the `preserved-thinking-migration` subcommand:
137
138```text wrap
139/claude-api preserved-thinking-migration
140```
141
142The skill works through five steps:
143
1441. **Scope:** Checks whether your integration is affected, then sends three test requests to confirm that the API reports a deliberate edit.
1452. **Find:** Captures a few of your own multi-turn sessions, diffs consecutive requests, and lists each place your code edits the prefix.
1463. **Measure:** Replays the sessions with `"drop_block"` and reports which conversations lose `thinking` blocks, at which turn, and why.
1474. **Fix:** Replaces one edit at a time with the matching pattern in [Make changes without editing the prefix](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#replace-prefix-edits), and measures again after each change.
1485. **Report:** Summarizes each cause, the change made for it, and the counts before and after.
149
150If your project has an eval, the skill also compares scores and token usage with and without dropped `thinking` blocks.
151
152The test requests and replays are real Messages API requests, billed at the model's normal rates. Each replay requests only a few output tokens and sets `tool_choice` to `none`, so no tool runs. The skill states the cost and asks for your approval before it sends them.
132153
133154## Setting up a Managed Agent
134155
agents-and-tools/mcp-connector Changed · +9 / -11 lines
from line 1188
11881188
11891189## Client-side MCP helpers
11901190
1191If you manage your own MCP client connection (for example, with local stdio servers, MCP prompts, or MCP resources), the SDKs provide helper functions that convert between MCP types and Claude API types. This eliminates manual conversion code when using an MCP SDK for your language (for example, the [TypeScript MCP SDK](https://github.com/modelcontextprotocol/typescript-sdk)) alongside the Anthropic SDK.
1191If you manage your own MCP client connection (for example, with local stdio servers, MCP prompts, or MCP resources), the SDK provides helper functions that convert between MCP types and Claude API types. This eliminates manual conversion code when using an MCP SDK for your language (for example, the [TypeScript MCP SDK](https://github.com/modelcontextprotocol/typescript-sdk)) alongside the Anthropic SDK.
11921192
11931193<Note>
11941194 Use the [`mcp_servers` API parameter](https://platform.claude.com/docs/en/agents-and-tools/mcp-connector#using-the-mcp-connector-in-the-messages-api) when you have remote servers accessible by URL and only need tool support. Use the client-side helpers when you need local servers, prompts, resources, or more control over the connection with the base SDK.
from line 1326
13261326 ```
13271327</CodeGroup>
13281328
1329Helper names and exact signatures follow each language's conventions; this table shows the TypeScript forms:
1329| Helper | Description |
1330| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
1331| `async_mcp_tool(tool, mcp_client)` (python; typescript: `mcpTools(tools, mcpClient)`; csharp: `BetaMcp.ListToolsAsync(mcpClient)`; go: `mcp.NewBetaTools(tools, session)`; java: `BetaMcp.mcpTools(tools, mcpClient)`; php: `BetaMcp::tools($tools, $mcp)`; ruby: `Anthropic::Mcp.tools(tools, mcp_client)`) | Converts MCP tools to Claude API tools for use with `client.beta.messages.tool_runner()` (python, ruby; typescript: `client.beta.messages.toolRunner()`; java: `client.beta().messages().toolRunner()`; php: `$client->beta->messages->toolRunner()`; csharp: `client.Beta.Messages.ToolRunner()`; go: `client.Beta.Messages.NewToolRunner()`) |
1332| `mcp_message(message)` (python; typescript: `mcpMessages(messages)`; csharp: `BetaMcp.Messages(messages)`; go: `mcp.ToMessage(message)`; java: `BetaMcp.mcpMessages(messages)`; php: `BetaMcp::message($message)`; ruby: `Anthropic::Mcp.message(message)`) | Converts MCP prompt messages to Claude API message format |
1333| `mcp_resource_to_content(resource)` (python; typescript: `mcpResourceToContent(resource)`; csharp: `BetaMcp.ResourceToContent(resource)`; go: `mcp.ResourceToBlock(resource)`; java: `BetaMcp.mcpResourceContents(resource)`; php: `BetaMcp::resourceToContent($resource)`; ruby: `Anthropic::Mcp.resource_to_contents(resource)`) | Converts an MCP resource to a Claude API content block |
1334| `mcp_resource_to_file(resource)` (python; typescript: `mcpResourceToFile(resource)`; csharp: `BetaMcp.ResourceToFile(resource)`; go: `mcp.ResourceToFile(resource)`; java: `BetaMcp.mcpResourceFiles(resource)`; php: `BetaMcp::resourceToFile($resource)`; ruby: `Anthropic::Mcp.resource_to_files(resource)`) | Converts an MCP resource to a file object for upload |
13301335
1331| Helper | Description |
1332| -------------------------------- | --------------------------------------------------------------------------------------- |
1333| `mcpTools(tools, mcpClient)` | Converts MCP tools to Claude API tools for use with `client.beta.messages.toolRunner()` |
1334| `mcpMessages(messages)` | Converts MCP prompt messages to Claude API message format |
1335| `mcpResourceToContent(resource)` | Converts an MCP resource to a Claude API content block |
1336| `mcpResourceToFile(resource)` | Converts an MCP resource to a file object for upload |
1337
13381336### Use MCP tools
13391337
13401338Convert MCP tools for use with the SDK's [tool runner](https://platform.claude.com/docs/en/agents-and-tools/tool-use/tool-runner), which handles tool execution automatically:
from line 2043
20452043
20462044## Deprecated version: mcp-client-2025-04-04
20472045
2048<Note type="warning">
2046<Warning>
20492047 This version is deprecated. Migrate to `mcp-client-2025-11-20` using the preceding [migration guide](https://platform.claude.com/docs/en/agents-and-tools/mcp-connector#migration-guide).
2050</Note>
2048</Warning>
20512049
20522050The previous version of the MCP connector included tool configuration directly in the MCP server definition:
20532051
agents-and-tools/tool-use/browser-use-tool Changed · +4 / -4 lines
from line 285
285285Browser use runs as an agent loop in your application: Claude returns member tool calls, your executor runs them against the browser, and you return the results until Claude answers in text.
286286
287287<Steps>
288 <Step title="Provide Claude with the browser use tool and a user prompt" icon="tool">
288 <Step title="Provide Claude with the browser use tool and a user prompt">
289289 * Add the `browser_toolset_20260801` entry, and optionally other tools, to your API request.
290290 * Include a user prompt that calls for working with webpages, for example, "Open example.com/docs and tell me how to get started."
291291 </Step>
292292
293 <Step title="Claude responds with member tool calls" icon="wrench">
293 <Step title="Claude responds with member tool calls">
294294 * Claude returns one or more `tool_use` blocks in a single assistant turn; several in one turn form a batch action, for example, `left_click`, then `type`, then `key`.
295295 * Each block's `name` is the member name, each carries `"toolset_name": "browser"`, and `input` holds only that member's parameters, with no `action` field. The response's `stop_reason` is `tool_use`.
296296 </Step>
297297
298 <Step title="Run the calls in order and return results" icon="browser">
298 <Step title="Run the calls in order and return results">
299299 * Iterate every `tool_use` block in `response.content` (don't assume there's exactly one) and run them sequentially, in the order they appear, because later calls usually depend on earlier ones.
300300 * Return one `tool_result` per block in a new `user` message, matched by `tool_use_id`, and echo `"toolset_name": "browser"` on each. Every call must be answered or the next request is rejected.
301301 * If a call fails, return `is_error: true` with a text description for that block, then apply the halt rule in [Batch actions](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-tool#batch-actions) to every later block in the turn.
302302 </Step>
303303
304 <Step title="Claude continues until the task is complete" icon="arrows-clockwise">
304 <Step title="Claude continues until the task is complete">
305305 * Claude reads the results (page text, accessibility trees, screenshots, tab state) and, if it needs more, returns further member calls, which takes you back to step 3.
306306 * Otherwise, it returns a text response to the user.
307307 </Step>
agents-and-tools/tool-use/computer-use-tool Changed · +6 / -6 lines
from line 317
317317## How computer use works
318318
319319<Steps>
320 <Step title="Provide Claude with the computer use tool and a user prompt" icon="tool">
320 <Step title="Provide Claude with the computer use tool and a user prompt">
321321 * Add the computer use toolset (and optionally other tools) to the `tools` array of your API request.
322322 * Include a user prompt that requires desktop interaction, for example, "Save a picture of a cat to my desktop."
323323 </Step>
324324
325 <Step title="Claude responds with member tool calls" icon="wrench">
325 <Step title="Claude responds with member tool calls">
326326 * Claude assesses whether acting on the desktop can help with the user's query.
327327 * If so, Claude responds with one or more member `tool_use` blocks, such as `screenshot`, `left_click`, or `type`, each carrying `"toolset_name": "computer"`. A response with several of these blocks is a [batch action](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool#batch-actions).
328328 * The API response has a `stop_reason` of `tool_use`, signaling a tool use request.
329329 </Step>
330330
331 <Step title="Run the calls in order and return results" icon="computer">
331 <Step title="Run the calls in order and return results">
332332 * Iterate over every `tool_use` block in the response, in order. For each one, dispatch on the member `name` together with `toolset_name`, and perform that action with the block's `input` on your container or virtual machine.
333333 * Continue the conversation with a new `user` message that contains one `tool_result` block per `tool_use` block, matched by `tool_use_id` and each echoing `"toolset_name": "computer"`. Return an image for `screenshot` and `zoom`; a short text such as `OK` is enough for the other actions.
334334 * If an action fails, return `is_error: true` for that block and answer the rest of the batch as described in [Batch actions](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool#batch-actions).
335335 </Step>
336336
337 <Step title="Claude continues until the task is complete" icon="arrows-clockwise">
337 <Step title="Claude continues until the task is complete">
338338 * Claude analyzes the tool results to determine if more actions are needed or the task has been completed.
339339 * If Claude determines more actions are needed, it responds with another `tool_use` `stop_reason` and you should return to step 3.
340340 * Otherwise, it returns a text response to the user.
from line 2118
21182118
21192119Upgrading from `computer_20251124` to the toolset is optional: the models listed for `computer_20251124` under [Earlier tool versions](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool#earlier-tool-versions) keep accepting it with its beta header, so an existing integration keeps working until you change it. Claude 5.5 and later models are the exception on the Claude API and Google Cloud: there they accept only the toolset. Upgrade an integration before you move it to one of them. On Amazon Bedrock, Claude Opus 5.5 and Claude Sonnet 5.5 keep accepting `computer_20251124`. To upgrade, make the following changes together:
21202120
21211. **Remove the beta header.** Drop `anthropic-beta: computer-use-2025-11-24` from your requests. In the SDKs, remove the `betas` parameter and call the Messages API through the standard client rather than the beta namespace.
21211. **Remove the beta header.** Drop `anthropic-beta: computer-use-2025-11-24` from your requests. With the SDK, remove `betas` (python, typescript, php, ruby; csharp, go: `Betas`; java: `.addBeta()`) and call the Messages API through the standard client rather than the beta namespace.
212221222. **Change the `tools` entry.** Set `type` to `computer_toolset_20260801` and delete `name`, `display_width_px`, `display_height_px`, `display_number`, and `enable_zoom`. The toolset rejects each of these fields.
212321233. **Choose whether to keep zoom enabled.** Zoom is enabled by default on the toolset, whereas `enable_zoom` defaults to `false`. If your environment doesn't implement zoom, add `"configs": {"zoom": {"enabled": false}}` to keep the previous behavior; otherwise implement it (see [Available actions](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool#available-actions)).
212421244. **Handle every block in a turn.** Update your agent loop to iterate over every `tool_use` block in a response rather than reading only the first, and to dispatch on the block's `name` together with `toolset_name` instead of on `input.action`. Member inputs no longer contain an `action` field; the remaining fields are unchanged.
from line 2174
21742174
21752175## Earlier tool versions
21762176
2177Two earlier versions of the computer use tool remain available in beta for existing integrations, for models that don't support the toolset, and on platforms where the toolset isn't currently available. Each requires its [beta header](https://platform.claude.com/docs/en/api/beta-headers) on every request, and their parameters are documented in the [beta Messages API reference](https://platform.claude.com/docs/en/api/beta/messages/create). In the SDKs, pass the header through the `betas` parameter and use the beta namespace; only the computer use tool needs the header, not the bash or text editor tools in the same request.
2177Two earlier versions of the computer use tool remain available in beta for existing integrations, for models that don't support the toolset, and on platforms where the toolset isn't currently available. Each requires its [beta header](https://platform.claude.com/docs/en/api/beta-headers) on every request, and their parameters are documented in the [beta Messages API reference](https://platform.claude.com/docs/en/api/beta/messages/create). With the SDK, pass the header through `betas` (python, typescript, php, ruby; csharp, go: `Betas`; java: `.addBeta()`) and use the beta namespace; only the computer use tool needs the header, not the bash or text editor tools in the same request.
21782178
21792179| Tool version | Beta header | Use with | Parameters |
21802180| ------------------- | ------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------- |
agents-and-tools/tool-use/tool-runner Changed · +61 / -5 lines
from line 1106
11061106
11071107### Automatic context management
11081108
1109For long-running agentic tasks, the TypeScript and Ruby tool runners support automatic [compaction](https://platform.claude.com/docs/en/build-with-claude/context-editing#client-side-compaction-sdk), which generates summaries when token usage exceeds a threshold so the conversation can continue beyond context window limits. Both SDKs have deprecated this client-side option in favor of [server-side compaction](https://platform.claude.com/docs/en/build-with-claude/compaction-threshold), which works with every SDK's tool runner through the `context_management` request parameter. The Python SDK (v1.0 and later) and the Go, Java, C#, and PHP tool runners don't include client-side compaction. The Python, TypeScript, C#, Go, Java, PHP, and Ruby tool runners have a `compact_before_next_turn()` (typescript, java, php: `compactBeforeNextTurn()`; csharp, go: `CompactBeforeNextTurn()`) helper for on-demand compaction. See [Compact in a loop](https://platform.claude.com/docs/en/build-with-claude/compaction-on-demand#compact-in-a-loop). Use it or a `context_management` compaction edit on a runner, not both.
1109For long-running agentic tasks, the TypeScript and Ruby tool runners support automatic [compaction](https://platform.claude.com/docs/en/build-with-claude/context-editing#client-side-compaction-sdk), which generates summaries when token usage exceeds a threshold so the conversation can continue beyond context window limits. Both SDKs have deprecated this client-side option in favor of [server-side compaction](https://platform.claude.com/docs/en/build-with-claude/compaction-threshold), which works with every SDK's tool runner through the `context_management` request parameter. The Python SDK (v1.0 and later) and the Go, Java, C#, and PHP tool runners don't include client-side compaction. The tool runner has a `compact_before_next_turn()` (typescript, java, php: `compactBeforeNextTurn()`; csharp, go: `CompactBeforeNextTurn()`) helper for on-demand compaction. See [Compact in a loop](https://platform.claude.com/docs/en/build-with-claude/compaction-on-demand#compact-in-a-loop). Use it or a `context_management` compaction edit on a runner, not both.
11101110
11111111### Debugging tool execution
11121112
from line 1128
11281128
11291129By default, tool errors are passed back to Claude, which can then respond appropriately. However, you might want to detect errors and handle them differently, for example, to stop execution early or implement custom error handling.
11301130
1131In the Python and TypeScript SDKs, use the tool response method (`generate_tool_call_response()` in Python, `generateToolResponse()` in TypeScript) to intercept tool results and check for errors before they're sent to Claude. The other SDKs don't expose that hook. Their tabs describe the closest alternative:
1132
11331131<Tabs exclude="shell">
11341132 <Tab title="Python">
11351133 ```python
from line 1157
11591157 # Process the message normally
11601158 print(message.content)
11611159 ```
1160
1161 <Note>
1162 Call
1163
1164 `runner.generate_tool_call_response()`
1165
1166 in the loop to get the tool results and check them for errors before the runner sends them to Claude.
1167 </Note>
11621168 </Tab>
11631169
11641170 <Tab title="TypeScript">
from line 1198
11921198 console.log(message.content);
11931199 }
11941200 ```
1201
1202 <Note>
1203 Call
1204
1205 `runner.generateToolResponse()`
1206
1207 in the loop to get the tool results and check them for errors before the runner sends them to Claude.
1208 </Note>
11951209 </Tab>
11961210
11971211 <Tab title="C#">
from line 1312
12981312 break if message.stop_reason != :tool_use
12991313 end
13001314 ```
1315
1316 <Note>
1317 The Ruby tool runner has no hook that returns the tool results. Once
1318
1319 `runner.next_message`
1320
1321 returns, they are the last entry in
1322
1323 `runner.params[:messages]`
1324
1325 , so check them there before the next request sends them to Claude.
1326 </Note>
13011327 </Tab>
13021328</Tabs>
13031329
from line 1331
13051331
13061332You can modify tool results before they're sent back to Claude. This is useful for adding metadata such as `cache_control` to enable [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) on tool results, or for transforming the tool output.
13071333
1308In the Python and TypeScript SDKs, use the tool response method to get the tool result, then modify it before the runner proceeds. Whether you explicitly append the modified result or mutate it in place depends on the SDK. See the code comments in each tab.
1309
13101334<Tabs exclude="shell">
13111335 <Tab title="Python">
13121336 ```python
from line 1364
13401364
13411365 print(message.content)
13421366 ```
1367
1368 <Note>
1369 Call
1370
1371 `runner.generate_tool_call_response()`
1372
1373 to get the tool result, modify it, and pass it to
1374
1375 `runner.append_messages()`
1376
1377 so the runner does not append the original.
1378 </Note>
13431379 </Tab>
13441380
13451381 <Tab title="TypeScript">
from line 1409
13731409 console.log(message.content);
13741410 }
13751411 ```
1412
1413 <Note>
1414 Call
1415
1416 `runner.generateToolResponse()`
1417
1418 to get the tool result, then modify it in place. The runner appends the modified result for you.
1419 </Note>
13761420 </Tab>
13771421
13781422 <Tab title="C#">
from line 1568
15241568 break if message.stop_reason != :tool_use
15251569 end
15261570 ```
1571
1572 <Note>
1573 The Ruby tool runner has no hook that returns the tool results. Once
1574
1575 `runner.next_message`
1576
1577 returns, they are the last entry in
1578
1579 `runner.params[:messages]`
1580
1581 , so modify them there before the next request sends them to Claude.
1582 </Note>
15271583 </Tab>
15281584</Tabs>
15291585
api/beta Changed · +18053 / -18539 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/beta-headers Changed · +8 / -8 lines
from line 1
11---
22title: Beta headers
33url: https://platform.claude.com/docs/en/api/beta-headers
4description: Access experimental features before they become part of the standard API with the `anthropic-beta` header or the SDKs' `betas` parameter.
4description: Access experimental features before they become part of the standard API with the `anthropic-beta` header or the SDK's `betas` parameter.
55---
66
77Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
88
99<Info>
10 Each [client SDK](https://platform.claude.com/docs/en/cli-sdks-libraries/overview) exposes a `beta` namespace for calling the API with beta features enabled.
10 The [client SDK](https://platform.claude.com/docs/en/cli-sdks-libraries/overview) exposes a `client.beta` (python, typescript, ruby; csharp, go: `client.Beta`; java: `client.beta()`; php: `$client->beta`) namespace for calling the API with beta features enabled.
1111</Info>
1212
1313## How to use beta headers
from line 24
2424
2525Each feature's documentation states the exact beta name to send. The [API overview](https://platform.claude.com/docs/en/api/overview) lists the APIs currently in beta.
2626
27The following examples show the same request with cURL, the `ant` CLI, and the SDKs, using the [context editing](https://platform.claude.com/docs/en/build-with-claude/context-editing) beta as the example. The SDKs take beta names in the `betas` parameter and send the `anthropic-beta` header for you:
27The following examples show the same request with cURL, the `ant` CLI, and the SDK, using the [context editing](https://platform.claude.com/docs/en/build-with-claude/context-editing) beta as the example. The SDK takes beta names through `betas` (python, typescript, php, ruby; csharp, go: `Betas`; java: `.addBeta()`) and sends the `anthropic-beta` header for you:
2828
2929<CodeGroup>
3030 ```bash cURL
from line 176
176176anthropic-beta: feature3
177177```
178178
179When using an SDK, list each feature in the `betas` parameter (for example, `betas=["feature1", "feature2"]`). With the CLI, pass a single `--beta` flag with the feature names separated by commas (for example, `--beta feature1,feature2`). You can also repeat the flag (for example, `--beta feature1 --beta feature2`).
179With the SDK, list each feature (for example, `betas=["feature1", "feature2"]` (python; typescript, ruby: `betas: ["feature1", "feature2"]`; php: `betas: ['feature1', 'feature2']`; csharp: `Betas = ["feature1", "feature2"]`; go: `Betas: []anthropic.AnthropicBeta{"feature1", "feature2"}`; java: `.addBeta("feature1").addBeta("feature2")`)). With the CLI, pass a single `--beta` flag with the feature names separated by commas (for example, `--beta feature1,feature2`). You can also repeat the flag (for example, `--beta feature1 --beta feature2`).
180180
181181### Endpoint-specific headers
182182
from line 188
188188| `/v1/tunnels` | `mcp-tunnels-2026-06-22` |
189189| `/v1/memory_stores` and sub-resources | `agent-memory-2026-07-22` |
190190
191The SDKs' `beta` namespaces add these headers automatically. Add them yourself only when making raw HTTP requests. See the [Managed Agents overview](https://platform.claude.com/docs/en/managed-agents/overview), [Using agent memory](https://platform.claude.com/docs/en/managed-agents/memory), and the [MCP tunnels reference](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/reference#tunnels-api) for details.
191The SDK's `client.beta` (python, typescript, ruby; csharp, go: `client.Beta`; java: `client.beta()`; php: `$client->beta`) namespace adds these headers automatically. Add them yourself only when making raw HTTP requests. See the [Managed Agents overview](https://platform.claude.com/docs/en/managed-agents/overview), [Using agent memory](https://platform.claude.com/docs/en/managed-agents/memory), and the [MCP tunnels reference](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/reference#tunnels-api) for details.
192192
193Endpoint-specific headers that apply to the same endpoint aren't always combinable. On memory store endpoints, `agent-memory-2026-07-22` replaces `managed-agents-2026-04-01`: sending both on the same request returns a `400` error. The client SDKs send the correct header for each endpoint automatically.
193Endpoint-specific headers that apply to the same endpoint aren't always combinable. On memory store endpoints, `agent-memory-2026-07-22` replaces `managed-agents-2026-04-01`: sending both on the same request returns a `400` error. The SDK sends the correct header for each endpoint automatically.
194194
195195### Version naming conventions
196196
from line 219
219219
220220<CardGroup cols={2}>
221221 <Card title="Errors" icon="info" href="https://platform.claude.com/docs/en/api/errors">
222 Understand the HTTP status codes, error response shape, and request IDs the Claude API returns, and handle errors with the SDKs' typed exceptions.
222 Understand the HTTP status codes, error response shape, and request IDs the Claude API returns, and handle errors with the SDK's typed exceptions.
223223 </Card>
224224
225225 <Card title="API overview" icon="compass" href="https://platform.claude.com/docs/en/api/overview">
api/beta/agents Changed · +3717 / -3819 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 19
1919
2020 - `string`
2121
22 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
23
24 - `"message-batches-2024-09-24"`
25
26 - `"prompt-caching-2024-07-31"`
27
28 - `"computer-use-2024-10-22"`
29
30 - `"computer-use-2025-01-24"`
31
32 - `"pdfs-2024-09-25"`
33
34 - `"token-counting-2024-11-01"`
35
36 - `"token-efficient-tools-2025-02-19"`
37
38 - `"output-128k-2025-02-19"`
39
40 - `"files-api-2025-04-14"`
41
42 - `"mcp-client-2025-04-04"`
43
44 - `"mcp-client-2025-11-20"`
45
46 - `"dev-full-thinking-2025-05-14"`
47
48 - `"interleaved-thinking-2025-05-14"`
49
50 - `"code-execution-2025-05-22"`
51
52 - `"extended-cache-ttl-2025-04-11"`
53
54 - `"context-1m-2025-08-07"`
55
56 - `"context-management-2025-06-27"`
57
58 - `"model-context-window-exceeded-2025-08-26"`
59
60 - `"skills-2025-10-02"`
61
62 - `"fast-mode-2026-02-01"`
63
64 - `"output-300k-2026-03-24"`
65
66 - `"user-profiles-2026-03-24"`
67
68 - `"user-profiles-2026-08-18"`
69
70 - `"user-profiles-2026-09-04"`
71
72 - `"advisor-tool-2026-03-01"`
73
74 - `"managed-agents-2026-04-01"`
75
76 - `"cache-diagnosis-2026-04-07"`
77
78 - `"dreaming-2026-04-21"`
79
80 - `"thinking-token-count-2026-05-13"`
81
82 - `"server-side-fallback-2026-06-01"`
83
84 - `"server-side-fallback-2026-07-01"`
85
86 - `"fallback-credit-2026-06-01"`
87
88 - `"fallback-credit-2026-07-01"`
89
90 - `"agent-memory-2026-07-22"`
91
92 - `"mid-conversation-tool-changes-2026-07-01"`
93
94 - `"compact-2026-01-12"`
95
96 - `"computer-use-2025-11-24"`
97
98 - `"mcp-tunnels-2026-06-22"`
99
100 - `"structured-outputs-2025-11-13"`
101
102 - `"task-budgets-2026-03-13"`
103
104 - `"thinking-display-updates-2026-08-18"`
105
106 - `"ce-user-management-2026-07-13"`
107
108 - `"mid-conversation-output-config-2026-07-01"`
109
110 - `"thinking-binding-controls-2026-08-01"`
111
112 - `"mid-conversation-system-clear-at-2026-08-21"`
113
114 - `"compact-2026-09-04"`
115
116 - `"inline-tools-2026-09-15"`
117
118 - `"mcp-client-2026-09-15"`
22 - `"message-batches-2024-09-24"`
23
24 - `"prompt-caching-2024-07-31"`
25
26 - `"computer-use-2024-10-22"`
27
28 - `"computer-use-2025-01-24"`
29
30 - `"pdfs-2024-09-25"`
31
32 - `"token-counting-2024-11-01"`
33
34 - `"token-efficient-tools-2025-02-19"`
35
36 - `"output-128k-2025-02-19"`
37
38 - `"files-api-2025-04-14"`
39
40 - `"mcp-client-2025-04-04"`
41
42 - `"mcp-client-2025-11-20"`
43
44 - `"dev-full-thinking-2025-05-14"`
45
46 - `"interleaved-thinking-2025-05-14"`
47
48 - `"code-execution-2025-05-22"`
49
50 - `"extended-cache-ttl-2025-04-11"`
51
52 - `"context-1m-2025-08-07"`
53
54 - `"context-management-2025-06-27"`
55
56 - `"model-context-window-exceeded-2025-08-26"`
57
58 - `"skills-2025-10-02"`
59
60 - `"fast-mode-2026-02-01"`
61
62 - `"output-300k-2026-03-24"`
63
64 - `"user-profiles-2026-03-24"`
65
66 - `"user-profiles-2026-08-18"`
67
68 - `"user-profiles-2026-09-04"`
69
70 - `"advisor-tool-2026-03-01"`
71
72 - `"managed-agents-2026-04-01"`
73
74 - `"cache-diagnosis-2026-04-07"`
75
76 - `"dreaming-2026-04-21"`
77
78 - `"thinking-token-count-2026-05-13"`
79
80 - `"server-side-fallback-2026-06-01"`
81
82 - `"server-side-fallback-2026-07-01"`
83
84 - `"fallback-credit-2026-06-01"`
85
86 - `"fallback-credit-2026-07-01"`
87
88 - `"agent-memory-2026-07-22"`
89
90 - `"mid-conversation-tool-changes-2026-07-01"`
91
92 - `"compact-2026-01-12"`
93
94 - `"computer-use-2025-11-24"`
95
96 - `"mcp-tunnels-2026-06-22"`
97
98 - `"structured-outputs-2025-11-13"`
99
100 - `"task-budgets-2026-03-13"`
101
102 - `"thinking-display-updates-2026-08-18"`
103
104 - `"ce-user-management-2026-07-13"`
105
106 - `"mid-conversation-output-config-2026-07-01"`
107
108 - `"thinking-binding-controls-2026-08-01"`
109
110 - `"mid-conversation-system-clear-at-2026-08-21"`
111
112 - `"compact-2026-09-04"`
113
114 - `"inline-tools-2026-09-15"`
115
116 - `"mcp-client-2026-09-15"`
119117
120118- `"anthropic-workspace-id": optional string`
121119
from line 133
135133
136134 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
137135
138 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
136 - `"claude-sonnet-5-5"`
137
138 Efficient model for coding and agents
139
140 - `"claude-opus-5-5"`
141
142 Powerful intelligence for coding, knowledge work, and long-running agents
143
144 - `"claude-fable-5-1"`
145
146 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
147
148 - `"claude-sonnet-5"`
149
150 Efficient model for coding and agents
151
152 - `"claude-fable-5"`
153
154 Next generation of intelligence for the hardest knowledge work and coding problems
155
156 - `"claude-opus-5"`
157
158 Powerful intelligence for long-running agents and coding
159
160 - `"claude-opus-4-8"`
161
162 Powerful intelligence for long-running agents and coding
163
164 - `"claude-opus-4-7"`
165
166 Powerful intelligence for long-running agents and coding
167
168 - `"claude-opus-4-6"`
169
170 Powerful intelligence for long-running agents and coding
171
172 - `"claude-sonnet-4-6"`
173
174 Best combination of speed and intelligence
175
176 - `"claude-haiku-4-5"`
177
178 Fastest model with near-frontier intelligence
179
180 - `"claude-haiku-4-5-20251001"`
181
182 Fastest model with near-frontier intelligence
183
184 - `"claude-opus-4-5"`
185
186 Powerful intelligence for long-running agents and coding
187
188 - `"claude-opus-4-5-20251101"`
189
190 Powerful intelligence for long-running agents and coding
191
192 - `"claude-sonnet-4-5"`
193
194 High-performance model for agents and coding
195
196 - `"claude-sonnet-4-5-20250929"`
197
198 High-performance model for agents and coding
199
200 - `string`
201
202 - `BetaManagedAgentsModelConfigParams object`
203
204 An object that defines additional configuration control over model use
205
206 - `id: BetaManagedAgentsModel`
139207
140208 The model that will power your agent.
141209
142210 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
143211
212 - `effort: optional "low" or "medium" or "high" or 2 more or BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or 3 more or null`
213
214 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
215
216 - `"low"`
217
218 Low effort. Favors latency over reasoning depth.
219
220 - `"medium"`
221
222 Medium effort. Balances latency and reasoning depth.
223
224 - `"high"`
225
226 High effort. Favors reasoning depth.
227
228 - `"xhigh"`
229
230 Extra-high effort. Not all models accept this level.
231
232 - `"max"`
233
234 Maximum effort. Favors reasoning depth over latency.
235
236 - `BetaManagedAgentsEffortLow object`
237
238 Low effort. Favors latency over reasoning depth.
239
240 - `type: "low"`
241
242 - `BetaManagedAgentsEffortMedium object`
243
244 Medium effort. Balances latency and reasoning depth.
245
246 - `type: "medium"`
247
248 - `BetaManagedAgentsEffortHigh object`
249
250 High effort. Favors reasoning depth.
251
252 - `type: "high"`
253
254 - `BetaManagedAgentsEffortXhigh object`
255
256 Extra-high effort. Not all models accept this level.
257
258 - `type: "xhigh"`
259
260 - `BetaManagedAgentsEffortMax object`
261
262 Maximum effort. Favors reasoning depth over latency.
263
264 - `type: "max"`
265
266 - `inference_geo: optional string or null`
267
268 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo. On update, `model` is whole-object replacement — omitting inference_geo clears it.
269
270 - `speed: optional "standard" or "fast" or null`
271
272 Inference speed mode. Defaults to `standard`.
273
274 - `"standard"`
275
276 - `"fast"`
277
278- `name: string`
279
280 Human-readable name for the agent.
281
282 minLength: 1, maxLength: 256
283
284- `description: optional string or null`
285
286 Description of what the agent does.
287
288 maxLength: 2048
289
290- `mcp_servers: optional array of BetaManagedAgentsURLMCPServerParams`
291
292 MCP servers this agent connects to. Maximum 20. Names must be unique within the array. Every server must be referenced by an `mcp_toolset` in `tools`; unreferenced servers are rejected. See the [MCP connector guide](https://platform.claude.com/docs/en/managed-agents/mcp-connector).
293
294 - `type: "url"`
295
296 - `name: string`
297
298 Unique name for this server, referenced by mcp_toolset configurations. 1-255 characters.
299
300 minLength: 1, maxLength: 255
301
302 - `url: string`
303
304 Endpoint URL for the MCP server.
305
306 maxLength: 2048
307
308- `metadata: optional map[string]`
309
310 Arbitrary key-value metadata. Maximum 16 pairs, keys up to 64 chars, values up to 512 chars.
311
312- `multiagent: optional BetaManagedAgentsMultiagentParams or null`
313
314 Multiagent orchestration configuration. Currently supports the `coordinator` topology with a roster of 1-20 agents.
315
316 - `type: "coordinator"`
317
318 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
319
320 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
321
322 - `string`
323
324 - `BetaManagedAgentsAgentParams object`
325
326 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
327
328 - `type: "agent"`
329
330 - `id: string`
331
332 The `agent` ID.
333
334 minLength: 1, maxLength: 128
335
336 - `version: optional number`
337
338 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
339
340 format: int32
341
342 - `BetaManagedAgentsMultiagentSelfParams object`
343
344 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
345
346 - `type: "self"`
347
348 - `BetaManagedAgentsAdvisorParams object`
349
350 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
351
352 - `type: "advisor"`
353
354 - `model: string`
355
356 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
357
358 minLength: 1, maxLength: 256
359
360- `skills: optional array of BetaManagedAgentsSkillParams`
361
362 Skills available to the agent.
363
364 - `BetaManagedAgentsAnthropicSkillParams object`
365
366 An Anthropic-managed skill.
367
368 - `type: "anthropic"`
369
370 - `skill_id: string`
371
372 Identifier of the Anthropic skill (e.g., "xlsx").
373
374 minLength: 1, maxLength: 64
375
376 - `version: optional string or null`
377
378 Version to pin. Defaults to latest if omitted.
379
380 minLength: 1, maxLength: 64
381
382 - `BetaManagedAgentsCustomSkillParams object`
383
384 A user-created custom skill.
385
386 - `type: "custom"`
387
388 - `skill_id: string`
389
390 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
391
392 minLength: 1, maxLength: 64
393
394 - `version: optional string or null`
395
396 Version to pin. Defaults to latest if omitted.
397
398 minLength: 1, maxLength: 64
399
400- `system: optional string or null`
401
402 System prompt for the agent.
403
404 maxLength: 100000
405
406- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams`
407
408 Tool configurations available to the agent. Maximum of 128 tools across all toolsets allowed.
409
410 - `BetaManagedAgentsAgentToolset20260401Params object`
411
412 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
413
414 - `type: "agent_toolset_20260401"`
415
416 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
417
418 Per-tool configuration overrides.
419
420 - `BetaManagedAgentsBashToolConfigParams object`
421
422 Configuration override for the bash tool.
423
424 - `type: optional "bash"`
425
426 - `name: "bash"`
427
428 Must be "bash".
429
430 - `enabled: optional boolean or null`
431
432 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
433
434 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
435
436 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
437
438 - `BetaManagedAgentsAlwaysAllowPolicy object`
439
440 Tool calls are automatically approved without user confirmation.
441
442 - `type: "always_allow"`
443
444 - `BetaManagedAgentsAlwaysAskPolicy object`
445
446 Tool calls require user confirmation before execution.
447
448 - `type: "always_ask"`
449
450 - `BetaManagedAgentsAutoPolicy object`
451
452 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
453
454 - `type: "auto"`
455
456 - `BetaManagedAgentsEditToolConfigParams object`
457
458 Configuration override for the edit tool.
459
460 - `type: optional "edit"`
461
462 - `name: "edit"`
463
464 Must be "edit".
465
466 - `enabled: optional boolean or null`
467
468 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
469
470 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
471
472 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
473
474 - `BetaManagedAgentsAlwaysAllowPolicy object`
475
476 Tool calls are automatically approved without user confirmation.
477
478 - `BetaManagedAgentsAlwaysAskPolicy object`
479
480 Tool calls require user confirmation before execution.
481
482 - `BetaManagedAgentsAutoPolicy object`
483
484 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
485
486 - `BetaManagedAgentsReadToolConfigParams object`
487
488 Configuration override for the read tool.
489
490 - `type: optional "read"`
491
492 - `name: "read"`
493
494 Must be "read".
495
496 - `enabled: optional boolean or null`
497
498 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
499
500 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
501
502 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
503
504 - `BetaManagedAgentsAlwaysAllowPolicy object`
505
506 Tool calls are automatically approved without user confirmation.
507
508 - `BetaManagedAgentsAlwaysAskPolicy object`
509
510 Tool calls require user confirmation before execution.
511
512 - `BetaManagedAgentsAutoPolicy object`
513
514 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
515
516 - `BetaManagedAgentsWriteToolConfigParams object`
517
518 Configuration override for the write tool.
519
520 - `type: optional "write"`
521
522 - `name: "write"`
523
524 Must be "write".
525
526 - `enabled: optional boolean or null`
527
528 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
529
530 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
531
532 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
533
534 - `BetaManagedAgentsAlwaysAllowPolicy object`
535
536 Tool calls are automatically approved without user confirmation.
537
538 - `BetaManagedAgentsAlwaysAskPolicy object`
539
540 Tool calls require user confirmation before execution.
541
542 - `BetaManagedAgentsAutoPolicy object`
543
544 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
545
546 - `BetaManagedAgentsGlobToolConfigParams object`
547
548 Configuration override for the glob tool.
549
550 - `type: optional "glob"`
551
552 - `name: "glob"`
553
554 Must be "glob".
555
556 - `enabled: optional boolean or null`
557
558 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
559
560 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
561
562 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
563
564 - `BetaManagedAgentsAlwaysAllowPolicy object`
565
566 Tool calls are automatically approved without user confirmation.
567
568 - `BetaManagedAgentsAlwaysAskPolicy object`
569
570 Tool calls require user confirmation before execution.
571
572 - `BetaManagedAgentsAutoPolicy object`
573
574 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
575
576 - `BetaManagedAgentsGrepToolConfigParams object`
577
578 Configuration override for the grep tool.
579
580 - `type: optional "grep"`
581
582 - `name: "grep"`
583
584 Must be "grep".
585
586 - `enabled: optional boolean or null`
587
588 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
589
590 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
591
592 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
593
594 - `BetaManagedAgentsAlwaysAllowPolicy object`
595
596 Tool calls are automatically approved without user confirmation.
597
598 - `BetaManagedAgentsAlwaysAskPolicy object`
599
600 Tool calls require user confirmation before execution.
601
602 - `BetaManagedAgentsAutoPolicy object`
603
604 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
605
606 - `BetaManagedAgentsWebFetchToolConfigParams object`
607
608 Configuration override for the web_fetch tool.
609
610 - `type: optional "web_fetch"`
611
612 - `name: "web_fetch"`
613
614 Must be "web_fetch".
615
616 - `allowed_domains: optional array of string`
617
618 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
619
620 - `blocked_domains: optional array of string`
621
622 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
623
624 - `enabled: optional boolean or null`
625
626 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
627
628 - `max_content_tokens: optional number or null`
629
630 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
631
632 format: int32
633
634 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
635
636 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
637
638 - `BetaManagedAgentsAlwaysAllowPolicy object`
639
640 Tool calls are automatically approved without user confirmation.
641
642 - `BetaManagedAgentsAlwaysAskPolicy object`
643
644 Tool calls require user confirmation before execution.
645
646 - `BetaManagedAgentsAutoPolicy object`
647
648 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
649
650 - `BetaManagedAgentsWebSearchToolConfigParams object`
651
652 Configuration override for the web_search tool.
653
654 - `type: optional "web_search"`
655
656 - `name: "web_search"`
657
658 Must be "web_search".
659
660 - `allowed_domains: optional array of string`
661
662 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
663
664 - `blocked_domains: optional array of string`
665
666 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
667
668 - `enabled: optional boolean or null`
669
670 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
671
672 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
673
674 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
675
676 - `BetaManagedAgentsAlwaysAllowPolicy object`
677
678 Tool calls are automatically approved without user confirmation.
679
680 - `BetaManagedAgentsAlwaysAskPolicy object`
681
682 Tool calls require user confirmation before execution.
683
684 - `BetaManagedAgentsAutoPolicy object`
685
686 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
687
688 - `user_location: optional BetaManagedAgentsUserLocation or null`
689
690 Approximate user location for search result localization.
691
692 - `type: "approximate"`
693
694 Location precision. Only "approximate" is supported.
695
696 - `city: optional string or null`
697
698 City name.
699
700 minLength: 1, maxLength: 255
701
702 - `country: optional string or null`
703
704 Two-letter ISO 3166-1 country code, uppercase.
705
706 - `region: optional string or null`
707
708 Region or state name.
709
710 minLength: 1, maxLength: 255
711
712 - `timezone: optional string or null`
713
714 IANA timezone identifier, e.g. "America/Los_Angeles".
715
716 minLength: 1, maxLength: 255
717
718 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
719
720 Default configuration applied to all tools in this set.
721
722 - `enabled: optional boolean or null`
723
724 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
725
726 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
727
728 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
729
730 - `BetaManagedAgentsAlwaysAllowPolicy object`
731
732 Tool calls are automatically approved without user confirmation.
733
734 - `BetaManagedAgentsAlwaysAskPolicy object`
735
736 Tool calls require user confirmation before execution.
737
738 - `BetaManagedAgentsAutoPolicy object`
739
740 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
741
742 - `BetaManagedAgentsMCPToolsetParams object`
743
744 Configuration for tools from an MCP server defined in `mcp_servers`.
745
746 - `type: "mcp_toolset"`
747
748 - `mcp_server_name: string`
749
750 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
751
752 minLength: 1, maxLength: 255
753
754 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
755
756 Per-tool configuration overrides.
757
758 - `name: string`
759
760 Name of the MCP tool to configure. 1-128 characters.
761
762 minLength: 1, maxLength: 128
763
764 - `enabled: optional boolean or null`
765
766 Whether this tool is enabled. Overrides the `default_config` setting.
767
768 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
769
770 Permission policy for this tool. Overrides the `default_config` setting.
771
772 - `BetaManagedAgentsAlwaysAllowPolicy object`
773
774 Tool calls are automatically approved without user confirmation.
775
776 - `BetaManagedAgentsAlwaysAskPolicy object`
777
778 Tool calls require user confirmation before execution.
779
780 - `BetaManagedAgentsAutoPolicy object`
781
782 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
783
784 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
785
786 Default configuration for all tools from this server.
787
788 - `enabled: optional boolean or null`
789
790 Whether tools are enabled by default. Defaults to true if not specified.
791
792 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
793
794 Default permission policy for tools from this server.
795
796 - `BetaManagedAgentsAlwaysAllowPolicy object`
797
798 Tool calls are automatically approved without user confirmation.
799
800 - `BetaManagedAgentsAlwaysAskPolicy object`
801
802 Tool calls require user confirmation before execution.
803
804 - `BetaManagedAgentsAutoPolicy object`
805
806 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
807
808 - `BetaManagedAgentsCustomToolParams object`
809
810 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
811
812 - `type: "custom"`
813
814 - `description: string`
815
816 Description of what the tool does, shown to the agent to help it decide when to use the tool.
817
818 minLength: 1
819
820 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
821
822 JSON Schema defining the expected input parameters for the tool.
823
824 - `type: "object"`
825
826 - `properties: optional map[unknown] or null`
827
828 - `required: optional array of string or null`
829
830 - `name: string`
831
832 Unique name for the tool. 1-128 characters; letters, digits, underscores, and hyphens.
833
834 minLength: 1, maxLength: 128
835
836### Returns
837
838- `BetaManagedAgentsAgent object`
839
840 A Managed Agents `agent`.
841
842 - `type: "agent"`
843
844 - `id: string`
845
846 - `archived_at: string or null`
847
848 When the agent was archived. Null if not archived.
849
850 format: date-time
851
852 - `created_at: string`
853
854 A timestamp in RFC 3339 format
855
856 format: date-time
857
858 - `description: string or null`
859
860 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
861
862 - `type: "url"`
863
864 - `name: string`
865
866 - `url: string`
867
868 - `metadata: map[string]`
869
870 - `model: BetaManagedAgentsModelConfig`
871
872 Model identifier and configuration.
873
874 - `id: BetaManagedAgentsModel`
875
876 The model that will power your agent.
877
878 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
879
144880 - `"claude-sonnet-5-5"`
145881
146882 Efficient model for coding and agents
from line 940
204940 - `"claude-sonnet-4-5-20250929"`
205941
206942 High-performance model for agents and coding
207
208 - `string`
209
210 - `BetaManagedAgentsModelConfigParams object`
211
212 An object that defines additional configuration control over model use
213
214 - `id: BetaManagedAgentsModel`
215
216 The model that will power your agent.
217
218 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
219
220 - `effort: optional "low" or "medium" or "high" or 2 more or BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or 3 more or null`
221
222 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
223
224 - `BetaManagedAgentsEffortLevel = "low" or "medium" or "high" or 2 more`
225
226 How hard Claude works on each turn. Higher levels favor reasoning depth over latency. Not all models accept every level; invalid combinations are rejected at create time.
227
228 - `"low"`
229
230 Low effort. Favors latency over reasoning depth.
231
232 - `"medium"`
233
234 Medium effort. Balances latency and reasoning depth.
235
236 - `"high"`
237
238 High effort. Favors reasoning depth.
239
240 - `"xhigh"`
241
242 Extra-high effort. Not all models accept this level.
243
244 - `"max"`
245
246 Maximum effort. Favors reasoning depth over latency.
247
248 - `BetaManagedAgentsEffortLow object`
249
250 Low effort. Favors latency over reasoning depth.
251
252 - `type: "low"`
253
254 - `BetaManagedAgentsEffortMedium object`
255
256 Medium effort. Balances latency and reasoning depth.
257
258 - `type: "medium"`
259
260 - `BetaManagedAgentsEffortHigh object`
261
262 High effort. Favors reasoning depth.
263
264 - `type: "high"`
265
266 - `BetaManagedAgentsEffortXhigh object`
267
268 Extra-high effort. Not all models accept this level.
269
270 - `type: "xhigh"`
271
272 - `BetaManagedAgentsEffortMax object`
273
274 Maximum effort. Favors reasoning depth over latency.
275
276 - `type: "max"`
277
278 - `inference_geo: optional string or null`
279
280 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo. On update, `model` is whole-object replacement — omitting inference_geo clears it.
281
282 - `speed: optional "standard" or "fast" or null`
283
284 Inference speed mode. Defaults to `standard`.
285
286 - `"standard"`
287
288 - `"fast"`
289
290- `name: string`
291
292 Human-readable name for the agent.
293
294 minLength: 1, maxLength: 256
295
296- `description: optional string or null`
297
298 Description of what the agent does.
299
300 maxLength: 2048
301
302- `mcp_servers: optional array of BetaManagedAgentsURLMCPServerParams`
303
304 MCP servers this agent connects to. Maximum 20. Names must be unique within the array. Every server must be referenced by an `mcp_toolset` in `tools`; unreferenced servers are rejected. See the [MCP connector guide](https://platform.claude.com/docs/en/managed-agents/mcp-connector).
305
306 - `type: "url"`
307
308 - `name: string`
309
310 Unique name for this server, referenced by mcp_toolset configurations. 1-255 characters.
311
312 minLength: 1, maxLength: 255
313
314 - `url: string`
315
316 Endpoint URL for the MCP server.
317
318 maxLength: 2048
319
320- `metadata: optional map[string]`
321
322 Arbitrary key-value metadata. Maximum 16 pairs, keys up to 64 chars, values up to 512 chars.
323
324- `multiagent: optional BetaManagedAgentsMultiagentParams or null`
325
326 Multiagent orchestration configuration. Currently supports the `coordinator` topology with a roster of 1-20 agents.
327
328 - `type: "coordinator"`
329
330 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
331
332 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
333
334 - `string`
335
336 - `BetaManagedAgentsAgentParams object`
337
338 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
339
340 - `type: "agent"`
341
342 - `id: string`
343
344 The `agent` ID.
345
346 minLength: 1, maxLength: 128
347
348 - `version: optional number`
349
350 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
351
352 format: int32
353
354 - `BetaManagedAgentsMultiagentSelfParams object`
355
356 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
357
358 - `type: "self"`
359
360 - `BetaManagedAgentsAdvisorParams object`
361
362 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
363
364 - `type: "advisor"`
365
366 - `model: string`
367
368 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
369
370 minLength: 1, maxLength: 256
371
372- `skills: optional array of BetaManagedAgentsSkillParams`
373
374 Skills available to the agent.
375
376 - `BetaManagedAgentsAnthropicSkillParams object`
377
378 An Anthropic-managed skill.
379
380 - `type: "anthropic"`
381
382 - `skill_id: string`
383
384 Identifier of the Anthropic skill (e.g., "xlsx").
385
386 minLength: 1, maxLength: 64
387
388 - `version: optional string or null`
389
390 Version to pin. Defaults to latest if omitted.
391
392 minLength: 1, maxLength: 64
393
394 - `BetaManagedAgentsCustomSkillParams object`
395
396 A user-created custom skill.
397
398 - `type: "custom"`
399
400 - `skill_id: string`
401
402 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
403
404 minLength: 1, maxLength: 64
405
406 - `version: optional string or null`
407
408 Version to pin. Defaults to latest if omitted.
409
410 minLength: 1, maxLength: 64
411
412- `system: optional string or null`
413
414 System prompt for the agent.
415
416 maxLength: 100000
417
418- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams`
419
420 Tool configurations available to the agent. Maximum of 128 tools across all toolsets allowed.
421
422 - `BetaManagedAgentsAgentToolset20260401Params object`
423
424 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
425
426 - `type: "agent_toolset_20260401"`
427
428 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
429
430 Per-tool configuration overrides.
431
432 - `BetaManagedAgentsBashToolConfigParams object`
433
434 Configuration override for the bash tool.
435
436 - `type: optional "bash"`
437
438 - `name: "bash"`
439
440 Must be "bash".
441
442 - `enabled: optional boolean or null`
443
444 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
445
446 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
447
448 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
449
450 - `BetaManagedAgentsAlwaysAllowPolicy object`
451
452 Tool calls are automatically approved without user confirmation.
453
454 - `type: "always_allow"`
455
456 - `BetaManagedAgentsAlwaysAskPolicy object`
457
458 Tool calls require user confirmation before execution.
459
460 - `type: "always_ask"`
461
462 - `BetaManagedAgentsAutoPolicy object`
463
464 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
465
466 - `type: "auto"`
467
468 - `BetaManagedAgentsEditToolConfigParams object`
469
470 Configuration override for the edit tool.
471
472 - `type: optional "edit"`
473
474 - `name: "edit"`
475
476 Must be "edit".
477
478 - `enabled: optional boolean or null`
479
480 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
481
482 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
483
484 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
485
486 - `BetaManagedAgentsAlwaysAllowPolicy object`
487
488 Tool calls are automatically approved without user confirmation.
489
490 - `BetaManagedAgentsAlwaysAskPolicy object`
491
492 Tool calls require user confirmation before execution.
493
494 - `BetaManagedAgentsAutoPolicy object`
495
496 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
497
498 - `BetaManagedAgentsReadToolConfigParams object`
499
500 Configuration override for the read tool.
501
502 - `type: optional "read"`
503
504 - `name: "read"`
505
506 Must be "read".
507
508 - `enabled: optional boolean or null`
509
510 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
511
512 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
513
514 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
515
516 - `BetaManagedAgentsAlwaysAllowPolicy object`
517
518 Tool calls are automatically approved without user confirmation.
519
520 - `BetaManagedAgentsAlwaysAskPolicy object`
521
522 Tool calls require user confirmation before execution.
523
524 - `BetaManagedAgentsAutoPolicy object`
525
526 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
527
528 - `BetaManagedAgentsWriteToolConfigParams object`
529
530 Configuration override for the write tool.
531
532 - `type: optional "write"`
533
534 - `name: "write"`
535
536 Must be "write".
537
538 - `enabled: optional boolean or null`
539
540 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
541
542 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
543
544 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
545
546 - `BetaManagedAgentsAlwaysAllowPolicy object`
547
548 Tool calls are automatically approved without user confirmation.
549
550 - `BetaManagedAgentsAlwaysAskPolicy object`
551
552 Tool calls require user confirmation before execution.
553
554 - `BetaManagedAgentsAutoPolicy object`
555
556 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
557
558 - `BetaManagedAgentsGlobToolConfigParams object`
559
560 Configuration override for the glob tool.
561
562 - `type: optional "glob"`
563
564 - `name: "glob"`
565
566 Must be "glob".
567
568 - `enabled: optional boolean or null`
569
570 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
571
572 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
573
574 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
575
576 - `BetaManagedAgentsAlwaysAllowPolicy object`
577
578 Tool calls are automatically approved without user confirmation.
579
580 - `BetaManagedAgentsAlwaysAskPolicy object`
581
582 Tool calls require user confirmation before execution.
583
584 - `BetaManagedAgentsAutoPolicy object`
585
586 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
587
588 - `BetaManagedAgentsGrepToolConfigParams object`
589
590 Configuration override for the grep tool.
591
592 - `type: optional "grep"`
593
594 - `name: "grep"`
595
596 Must be "grep".
597
598 - `enabled: optional boolean or null`
599
600 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
601
602 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
603
604 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
605
606 - `BetaManagedAgentsAlwaysAllowPolicy object`
607
608 Tool calls are automatically approved without user confirmation.
609
610 - `BetaManagedAgentsAlwaysAskPolicy object`
611
612 Tool calls require user confirmation before execution.
613
614 - `BetaManagedAgentsAutoPolicy object`
615
616 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
617
618 - `BetaManagedAgentsWebFetchToolConfigParams object`
619
620 Configuration override for the web_fetch tool.
621
622 - `type: optional "web_fetch"`
623
624 - `name: "web_fetch"`
625
626 Must be "web_fetch".
627
628 - `allowed_domains: optional array of string`
629
630 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
631
632 - `blocked_domains: optional array of string`
633
634 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
635
636 - `enabled: optional boolean or null`
637
638 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
639
640 - `max_content_tokens: optional number or null`
641
642 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
643
644 format: int32
645
646 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
647
648 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
649
650 - `BetaManagedAgentsAlwaysAllowPolicy object`
651
652 Tool calls are automatically approved without user confirmation.
653
654 - `BetaManagedAgentsAlwaysAskPolicy object`
655
656 Tool calls require user confirmation before execution.
657
658 - `BetaManagedAgentsAutoPolicy object`
659
660 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
661
662 - `BetaManagedAgentsWebSearchToolConfigParams object`
663
664 Configuration override for the web_search tool.
665
666 - `type: optional "web_search"`
667
668 - `name: "web_search"`
669
670 Must be "web_search".
671
672 - `allowed_domains: optional array of string`
673
674 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
675
676 - `blocked_domains: optional array of string`
677
678 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
679
680 - `enabled: optional boolean or null`
681
682 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
683
684 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
685
686 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
687
688 - `BetaManagedAgentsAlwaysAllowPolicy object`
689
690 Tool calls are automatically approved without user confirmation.
691
692 - `BetaManagedAgentsAlwaysAskPolicy object`
693
694 Tool calls require user confirmation before execution.
695
696 - `BetaManagedAgentsAutoPolicy object`
697
698 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
699
700 - `user_location: optional BetaManagedAgentsUserLocation or null`
701
702 Approximate user location for search result localization.
703
704 - `type: "approximate"`
705
706 Location precision. Only "approximate" is supported.
707
708 - `city: optional string or null`
709
710 City name.
711
712 minLength: 1, maxLength: 255
713
714 - `country: optional string or null`
715
716 Two-letter ISO 3166-1 country code, uppercase.
717
718 - `region: optional string or null`
719
720 Region or state name.
721
722 minLength: 1, maxLength: 255
723
724 - `timezone: optional string or null`
725
726 IANA timezone identifier, e.g. "America/Los_Angeles".
727
728 minLength: 1, maxLength: 255
729
730 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
731
732 Default configuration applied to all tools in this set.
733
734 - `enabled: optional boolean or null`
735
736 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
737
738 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
739
740 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
741
742 - `BetaManagedAgentsAlwaysAllowPolicy object`
743
744 Tool calls are automatically approved without user confirmation.
745
746 - `BetaManagedAgentsAlwaysAskPolicy object`
747
748 Tool calls require user confirmation before execution.
749
750 - `BetaManagedAgentsAutoPolicy object`
751
752 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
753
754 - `BetaManagedAgentsMCPToolsetParams object`
755
756 Configuration for tools from an MCP server defined in `mcp_servers`.
757
758 - `type: "mcp_toolset"`
759
760 - `mcp_server_name: string`
761
762 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
763
764 minLength: 1, maxLength: 255
765
766 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
767
768 Per-tool configuration overrides.
769
770 - `name: string`
771
772 Name of the MCP tool to configure. 1-128 characters.
773
774 minLength: 1, maxLength: 128
775
776 - `enabled: optional boolean or null`
777
778 Whether this tool is enabled. Overrides the `default_config` setting.
779
780 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
781
782 Permission policy for this tool. Overrides the `default_config` setting.
783
784 - `BetaManagedAgentsAlwaysAllowPolicy object`
785
786 Tool calls are automatically approved without user confirmation.
787
788 - `BetaManagedAgentsAlwaysAskPolicy object`
789
790 Tool calls require user confirmation before execution.
791
792 - `BetaManagedAgentsAutoPolicy object`
793
794 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
795
796 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
797
798 Default configuration for all tools from this server.
799
800 - `enabled: optional boolean or null`
801
802 Whether tools are enabled by default. Defaults to true if not specified.
803
804 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
805
806 Default permission policy for tools from this server.
807
808 - `BetaManagedAgentsAlwaysAllowPolicy object`
809
810 Tool calls are automatically approved without user confirmation.
811
812 - `BetaManagedAgentsAlwaysAskPolicy object`
813
814 Tool calls require user confirmation before execution.
815
816 - `BetaManagedAgentsAutoPolicy object`
817
818 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
819
820 - `BetaManagedAgentsCustomToolParams object`
821
822 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
823
824 - `type: "custom"`
825
826 - `description: string`
827
828 Description of what the tool does, shown to the agent to help it decide when to use the tool.
829
830 minLength: 1
831
832 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
833
834 JSON Schema defining the expected input parameters for the tool.
835
836 - `type: "object"`
837
838 - `properties: optional map[unknown] or null`
839
840 - `required: optional array of string or null`
841
842 - `name: string`
843
844 Unique name for the tool. 1-128 characters; letters, digits, underscores, and hyphens.
845
846 minLength: 1, maxLength: 128
847
848### Returns
849
850- `BetaManagedAgentsAgent object`
851
852 A Managed Agents `agent`.
853
854 - `type: "agent"`
855
856 - `id: string`
857
858 - `archived_at: string or null`
859
860 When the agent was archived. Null if not archived.
861
862 format: date-time
863
864 - `created_at: string`
865
866 A timestamp in RFC 3339 format
867
868 format: date-time
869
870 - `description: string or null`
871
872 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
873
874 - `type: "url"`
875
876 - `name: string`
877
878 - `url: string`
879
880 - `metadata: map[string]`
881
882 - `model: BetaManagedAgentsModelConfig`
883
884 Model identifier and configuration.
885
886 - `id: BetaManagedAgentsModel`
887
888 The model that will power your agent.
889
890 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
891
892 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
893
894 The model that will power your agent.
895
896 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
897
898 - `"claude-sonnet-5-5"`
899
900 Efficient model for coding and agents
901
902 - `"claude-opus-5-5"`
903
904 Powerful intelligence for coding, knowledge work, and long-running agents
905
906 - `"claude-fable-5-1"`
907
908 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
909
910 - `"claude-sonnet-5"`
911
912 Efficient model for coding and agents
913
914 - `"claude-fable-5"`
915
916 Next generation of intelligence for the hardest knowledge work and coding problems
917
918 - `"claude-opus-5"`
919
920 Powerful intelligence for long-running agents and coding
921
922 - `"claude-opus-4-8"`
923
924 Powerful intelligence for long-running agents and coding
925
926 - `"claude-opus-4-7"`
927
928 Powerful intelligence for long-running agents and coding
929
930 - `"claude-opus-4-6"`
931
932 Powerful intelligence for long-running agents and coding
933
934 - `"claude-sonnet-4-6"`
935
936 Best combination of speed and intelligence
937
938 - `"claude-haiku-4-5"`
939
940 Fastest model with near-frontier intelligence
941
942 - `"claude-haiku-4-5-20251001"`
943
944 Fastest model with near-frontier intelligence
945
946 - `"claude-opus-4-5"`
947
948 Powerful intelligence for long-running agents and coding
949
950 - `"claude-opus-4-5-20251101"`
951
952 Powerful intelligence for long-running agents and coding
953
954 - `"claude-sonnet-4-5"`
955
956 High-performance model for agents and coding
957
958 - `"claude-sonnet-4-5-20250929"`
959
960 High-performance model for agents and coding
961943
962944 - `string`
963945
from line 1567
15851567
15861568 - `string`
15871569
1588 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
1589
1590 - `"message-batches-2024-09-24"`
1591
1592 - `"prompt-caching-2024-07-31"`
1593
1594 - `"computer-use-2024-10-22"`
1595
1596 - `"computer-use-2025-01-24"`
1597
1598 - `"pdfs-2024-09-25"`
1599
1600 - `"token-counting-2024-11-01"`
1601
1602 - `"token-efficient-tools-2025-02-19"`
1603
1604 - `"output-128k-2025-02-19"`
1605
1606 - `"files-api-2025-04-14"`
1607
1608 - `"mcp-client-2025-04-04"`
1609
1610 - `"mcp-client-2025-11-20"`
1611
1612 - `"dev-full-thinking-2025-05-14"`
1613
1614 - `"interleaved-thinking-2025-05-14"`
1615
1616 - `"code-execution-2025-05-22"`
1617
1618 - `"extended-cache-ttl-2025-04-11"`
1619
1620 - `"context-1m-2025-08-07"`
1621
1622 - `"context-management-2025-06-27"`
1623
1624 - `"model-context-window-exceeded-2025-08-26"`
1625
1626 - `"skills-2025-10-02"`
1627
1628 - `"fast-mode-2026-02-01"`
1629
1630 - `"output-300k-2026-03-24"`
1631
1632 - `"user-profiles-2026-03-24"`
1633
1634 - `"user-profiles-2026-08-18"`
1635
1636 - `"user-profiles-2026-09-04"`
1637
1638 - `"advisor-tool-2026-03-01"`
1639
1640 - `"managed-agents-2026-04-01"`
1641
1642 - `"cache-diagnosis-2026-04-07"`
1643
1644 - `"dreaming-2026-04-21"`
1645
1646 - `"thinking-token-count-2026-05-13"`
1647
1648 - `"server-side-fallback-2026-06-01"`
1649
1650 - `"server-side-fallback-2026-07-01"`
1651
1652 - `"fallback-credit-2026-06-01"`
1653
1654 - `"fallback-credit-2026-07-01"`
1655
1656 - `"agent-memory-2026-07-22"`
1657
1658 - `"mid-conversation-tool-changes-2026-07-01"`
1659
1660 - `"compact-2026-01-12"`
1661
1662 - `"computer-use-2025-11-24"`
1663
1664 - `"mcp-tunnels-2026-06-22"`
1665
1666 - `"structured-outputs-2025-11-13"`
1667
1668 - `"task-budgets-2026-03-13"`
1669
1670 - `"thinking-display-updates-2026-08-18"`
1671
1672 - `"ce-user-management-2026-07-13"`
1673
1674 - `"mid-conversation-output-config-2026-07-01"`
1675
1676 - `"thinking-binding-controls-2026-08-01"`
1677
1678 - `"mid-conversation-system-clear-at-2026-08-21"`
1679
1680 - `"compact-2026-09-04"`
1681
1682 - `"inline-tools-2026-09-15"`
1683
1684 - `"mcp-client-2026-09-15"`
1570 - `"message-batches-2024-09-24"`
1571
1572 - `"prompt-caching-2024-07-31"`
1573
1574 - `"computer-use-2024-10-22"`
1575
1576 - `"computer-use-2025-01-24"`
1577
1578 - `"pdfs-2024-09-25"`
1579
1580 - `"token-counting-2024-11-01"`
1581
1582 - `"token-efficient-tools-2025-02-19"`
1583
1584 - `"output-128k-2025-02-19"`
1585
1586 - `"files-api-2025-04-14"`
1587
1588 - `"mcp-client-2025-04-04"`
1589
1590 - `"mcp-client-2025-11-20"`
1591
1592 - `"dev-full-thinking-2025-05-14"`
1593
1594 - `"interleaved-thinking-2025-05-14"`
1595
1596 - `"code-execution-2025-05-22"`
1597
1598 - `"extended-cache-ttl-2025-04-11"`
1599
1600 - `"context-1m-2025-08-07"`
1601
1602 - `"context-management-2025-06-27"`
1603
1604 - `"model-context-window-exceeded-2025-08-26"`
1605
1606 - `"skills-2025-10-02"`
1607
1608 - `"fast-mode-2026-02-01"`
1609
1610 - `"output-300k-2026-03-24"`
1611
1612 - `"user-profiles-2026-03-24"`
1613
1614 - `"user-profiles-2026-08-18"`
1615
1616 - `"user-profiles-2026-09-04"`
1617
1618 - `"advisor-tool-2026-03-01"`
1619
1620 - `"managed-agents-2026-04-01"`
1621
1622 - `"cache-diagnosis-2026-04-07"`
1623
1624 - `"dreaming-2026-04-21"`
1625
1626 - `"thinking-token-count-2026-05-13"`
1627
1628 - `"server-side-fallback-2026-06-01"`
1629
1630 - `"server-side-fallback-2026-07-01"`
1631
1632 - `"fallback-credit-2026-06-01"`
1633
1634 - `"fallback-credit-2026-07-01"`
1635
1636 - `"agent-memory-2026-07-22"`
1637
1638 - `"mid-conversation-tool-changes-2026-07-01"`
1639
1640 - `"compact-2026-01-12"`
1641
1642 - `"computer-use-2025-11-24"`
1643
1644 - `"mcp-tunnels-2026-06-22"`
1645
1646 - `"structured-outputs-2025-11-13"`
1647
1648 - `"task-budgets-2026-03-13"`
1649
1650 - `"thinking-display-updates-2026-08-18"`
1651
1652 - `"ce-user-management-2026-07-13"`
1653
1654 - `"mid-conversation-output-config-2026-07-01"`
1655
1656 - `"thinking-binding-controls-2026-08-01"`
1657
1658 - `"mid-conversation-system-clear-at-2026-08-21"`
1659
1660 - `"compact-2026-09-04"`
1661
1662 - `"inline-tools-2026-09-15"`
1663
1664 - `"mcp-client-2026-09-15"`
16851665
16861666- `"anthropic-workspace-id": optional string`
16871667
from line 1713
17331713
17341714 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
17351715
1736 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
1737
1738 The model that will power your agent.
1739
1740 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
1741
1742 - `"claude-sonnet-5-5"`
1743
1744 Efficient model for coding and agents
1745
1746 - `"claude-opus-5-5"`
1747
1748 Powerful intelligence for coding, knowledge work, and long-running agents
1749
1750 - `"claude-fable-5-1"`
1751
1752 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
1753
1754 - `"claude-sonnet-5"`
1755
1756 Efficient model for coding and agents
1757
1758 - `"claude-fable-5"`
1759
1760 Next generation of intelligence for the hardest knowledge work and coding problems
1761
1762 - `"claude-opus-5"`
1763
1764 Powerful intelligence for long-running agents and coding
1765
1766 - `"claude-opus-4-8"`
1767
1768 Powerful intelligence for long-running agents and coding
1769
1770 - `"claude-opus-4-7"`
1771
1772 Powerful intelligence for long-running agents and coding
1773
1774 - `"claude-opus-4-6"`
1775
1776 Powerful intelligence for long-running agents and coding
1777
1778 - `"claude-sonnet-4-6"`
1779
1780 Best combination of speed and intelligence
1781
1782 - `"claude-haiku-4-5"`
1783
1784 Fastest model with near-frontier intelligence
1785
1786 - `"claude-haiku-4-5-20251001"`
1787
1788 Fastest model with near-frontier intelligence
1789
1790 - `"claude-opus-4-5"`
1791
1792 Powerful intelligence for long-running agents and coding
1793
1794 - `"claude-opus-4-5-20251101"`
1795
1796 Powerful intelligence for long-running agents and coding
1797
1798 - `"claude-sonnet-4-5"`
1799
1800 High-performance model for agents and coding
1801
1802 - `"claude-sonnet-4-5-20250929"`
1803
1804 High-performance model for agents and coding
1716 - `"claude-sonnet-5-5"`
1717
1718 Efficient model for coding and agents
1719
1720 - `"claude-opus-5-5"`
1721
1722 Powerful intelligence for coding, knowledge work, and long-running agents
1723
1724 - `"claude-fable-5-1"`
1725
1726 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
1727
1728 - `"claude-sonnet-5"`
1729
1730 Efficient model for coding and agents
1731
1732 - `"claude-fable-5"`
1733
1734 Next generation of intelligence for the hardest knowledge work and coding problems
1735
1736 - `"claude-opus-5"`
1737
1738 Powerful intelligence for long-running agents and coding
1739
1740 - `"claude-opus-4-8"`
1741
1742 Powerful intelligence for long-running agents and coding
1743
1744 - `"claude-opus-4-7"`
1745
1746 Powerful intelligence for long-running agents and coding
1747
1748 - `"claude-opus-4-6"`
1749
1750 Powerful intelligence for long-running agents and coding
1751
1752 - `"claude-sonnet-4-6"`
1753
1754 Best combination of speed and intelligence
1755
1756 - `"claude-haiku-4-5"`
1757
1758 Fastest model with near-frontier intelligence
1759
1760 - `"claude-haiku-4-5-20251001"`
1761
1762 Fastest model with near-frontier intelligence
1763
1764 - `"claude-opus-4-5"`
1765
1766 Powerful intelligence for long-running agents and coding
1767
1768 - `"claude-opus-4-5-20251101"`
1769
1770 Powerful intelligence for long-running agents and coding
1771
1772 - `"claude-sonnet-4-5"`
1773
1774 High-performance model for agents and coding
1775
1776 - `"claude-sonnet-4-5-20250929"`
1777
1778 High-performance model for agents and coding
18051779
18061780 - `string`
18071781
from line 2374
24002374
24012375 - `string`
24022376
2403 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
2404
2405 - `"message-batches-2024-09-24"`
2406
2407 - `"prompt-caching-2024-07-31"`
2408
2409 - `"computer-use-2024-10-22"`
2410
2411 - `"computer-use-2025-01-24"`
2412
2413 - `"pdfs-2024-09-25"`
2414
2415 - `"token-counting-2024-11-01"`
2416
2417 - `"token-efficient-tools-2025-02-19"`
2418
2419 - `"output-128k-2025-02-19"`
2420
2421 - `"files-api-2025-04-14"`
2422
2423 - `"mcp-client-2025-04-04"`
2424
2425 - `"mcp-client-2025-11-20"`
2426
2427 - `"dev-full-thinking-2025-05-14"`
2428
2429 - `"interleaved-thinking-2025-05-14"`
2430
2431 - `"code-execution-2025-05-22"`
2432
2433 - `"extended-cache-ttl-2025-04-11"`
2434
2435 - `"context-1m-2025-08-07"`
2436
2437 - `"context-management-2025-06-27"`
2438
2439 - `"model-context-window-exceeded-2025-08-26"`
2440
2441 - `"skills-2025-10-02"`
2442
2443 - `"fast-mode-2026-02-01"`
2444
2445 - `"output-300k-2026-03-24"`
2446
2447 - `"user-profiles-2026-03-24"`
2448
2449 - `"user-profiles-2026-08-18"`
2450
2451 - `"user-profiles-2026-09-04"`
2452
2453 - `"advisor-tool-2026-03-01"`
2454
2455 - `"managed-agents-2026-04-01"`
2456
2457 - `"cache-diagnosis-2026-04-07"`
2458
2459 - `"dreaming-2026-04-21"`
2460
2461 - `"thinking-token-count-2026-05-13"`
2462
2463 - `"server-side-fallback-2026-06-01"`
2464
2465 - `"server-side-fallback-2026-07-01"`
2466
2467 - `"fallback-credit-2026-06-01"`
2468
2469 - `"fallback-credit-2026-07-01"`
2470
2471 - `"agent-memory-2026-07-22"`
2472
2473 - `"mid-conversation-tool-changes-2026-07-01"`
2474
2475 - `"compact-2026-01-12"`
2476
2477 - `"computer-use-2025-11-24"`
2478
2479 - `"mcp-tunnels-2026-06-22"`
2480
2481 - `"structured-outputs-2025-11-13"`
2482
2483 - `"task-budgets-2026-03-13"`
2484
2485 - `"thinking-display-updates-2026-08-18"`
2486
2487 - `"ce-user-management-2026-07-13"`
2488
2489 - `"mid-conversation-output-config-2026-07-01"`
2490
2491 - `"thinking-binding-controls-2026-08-01"`
2492
2493 - `"mid-conversation-system-clear-at-2026-08-21"`
2494
2495 - `"compact-2026-09-04"`
2496
2497 - `"inline-tools-2026-09-15"`
2498
2499 - `"mcp-client-2026-09-15"`
2377 - `"message-batches-2024-09-24"`
2378
2379 - `"prompt-caching-2024-07-31"`
2380
2381 - `"computer-use-2024-10-22"`
2382
2383 - `"computer-use-2025-01-24"`
2384
2385 - `"pdfs-2024-09-25"`
2386
2387 - `"token-counting-2024-11-01"`
2388
2389 - `"token-efficient-tools-2025-02-19"`
2390
2391 - `"output-128k-2025-02-19"`
2392
2393 - `"files-api-2025-04-14"`
2394
2395 - `"mcp-client-2025-04-04"`
2396
2397 - `"mcp-client-2025-11-20"`
2398
2399 - `"dev-full-thinking-2025-05-14"`
2400
2401 - `"interleaved-thinking-2025-05-14"`
2402
2403 - `"code-execution-2025-05-22"`
2404
2405 - `"extended-cache-ttl-2025-04-11"`
2406
2407 - `"context-1m-2025-08-07"`
2408
2409 - `"context-management-2025-06-27"`
2410
2411 - `"model-context-window-exceeded-2025-08-26"`
2412
2413 - `"skills-2025-10-02"`
2414
2415 - `"fast-mode-2026-02-01"`
2416
2417 - `"output-300k-2026-03-24"`
2418
2419 - `"user-profiles-2026-03-24"`
2420
2421 - `"user-profiles-2026-08-18"`
2422
2423 - `"user-profiles-2026-09-04"`
2424
2425 - `"advisor-tool-2026-03-01"`
2426
2427 - `"managed-agents-2026-04-01"`
2428
2429 - `"cache-diagnosis-2026-04-07"`
2430
2431 - `"dreaming-2026-04-21"`
2432
2433 - `"thinking-token-count-2026-05-13"`
2434
2435 - `"server-side-fallback-2026-06-01"`
2436
2437 - `"server-side-fallback-2026-07-01"`
2438
2439 - `"fallback-credit-2026-06-01"`
2440
2441 - `"fallback-credit-2026-07-01"`
2442
2443 - `"agent-memory-2026-07-22"`
2444
2445 - `"mid-conversation-tool-changes-2026-07-01"`
2446
2447 - `"compact-2026-01-12"`
2448
2449 - `"computer-use-2025-11-24"`
2450
2451 - `"mcp-tunnels-2026-06-22"`
2452
2453 - `"structured-outputs-2025-11-13"`
2454
2455 - `"task-budgets-2026-03-13"`
2456
2457 - `"thinking-display-updates-2026-08-18"`
2458
2459 - `"ce-user-management-2026-07-13"`
2460
2461 - `"mid-conversation-output-config-2026-07-01"`
2462
2463 - `"thinking-binding-controls-2026-08-01"`
2464
2465 - `"mid-conversation-system-clear-at-2026-08-21"`
2466
2467 - `"compact-2026-09-04"`
2468
2469 - `"inline-tools-2026-09-15"`
2470
2471 - `"mcp-client-2026-09-15"`
25002472
25012473- `"anthropic-workspace-id": optional string`
25022474
from line 2520
25482520
25492521 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
25502522
2551 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
2552
2553 The model that will power your agent.
2554
2555 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
2556
2557 - `"claude-sonnet-5-5"`
2558
2559 Efficient model for coding and agents
2560
2561 - `"claude-opus-5-5"`
2562
2563 Powerful intelligence for coding, knowledge work, and long-running agents
2564
2565 - `"claude-fable-5-1"`
2566
2567 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
2568
2569 - `"claude-sonnet-5"`
2570
2571 Efficient model for coding and agents
2572
2573 - `"claude-fable-5"`
2574
2575 Next generation of intelligence for the hardest knowledge work and coding problems
2576
2577 - `"claude-opus-5"`
2578
2579 Powerful intelligence for long-running agents and coding
2580
2581 - `"claude-opus-4-8"`
2582
2583 Powerful intelligence for long-running agents and coding
2584
2585 - `"claude-opus-4-7"`
2586
2587 Powerful intelligence for long-running agents and coding
2588
2589 - `"claude-opus-4-6"`
2590
2591 Powerful intelligence for long-running agents and coding
2592
2593 - `"claude-sonnet-4-6"`
2594
2595 Best combination of speed and intelligence
2596
2597 - `"claude-haiku-4-5"`
2598
2599 Fastest model with near-frontier intelligence
2600
2601 - `"claude-haiku-4-5-20251001"`
2602
2603 Fastest model with near-frontier intelligence
2604
2605 - `"claude-opus-4-5"`
2606
2607 Powerful intelligence for long-running agents and coding
2608
2609 - `"claude-opus-4-5-20251101"`
2610
2611 Powerful intelligence for long-running agents and coding
2612
2613 - `"claude-sonnet-4-5"`
2614
2615 High-performance model for agents and coding
2616
2617 - `"claude-sonnet-4-5-20250929"`
2618
2619 High-performance model for agents and coding
2523 - `"claude-sonnet-5-5"`
2524
2525 Efficient model for coding and agents
2526
2527 - `"claude-opus-5-5"`
2528
2529 Powerful intelligence for coding, knowledge work, and long-running agents
2530
2531 - `"claude-fable-5-1"`
2532
2533 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
2534
2535 - `"claude-sonnet-5"`
2536
2537 Efficient model for coding and agents
2538
2539 - `"claude-fable-5"`
2540
2541 Next generation of intelligence for the hardest knowledge work and coding problems
2542
2543 - `"claude-opus-5"`
2544
2545 Powerful intelligence for long-running agents and coding
2546
2547 - `"claude-opus-4-8"`
2548
2549 Powerful intelligence for long-running agents and coding
2550
2551 - `"claude-opus-4-7"`
2552
2553 Powerful intelligence for long-running agents and coding
2554
2555 - `"claude-opus-4-6"`
2556
2557 Powerful intelligence for long-running agents and coding
2558
2559 - `"claude-sonnet-4-6"`
2560
2561 Best combination of speed and intelligence
2562
2563 - `"claude-haiku-4-5"`
2564
2565 Fastest model with near-frontier intelligence
2566
2567 - `"claude-haiku-4-5-20251001"`
2568
2569 Fastest model with near-frontier intelligence
2570
2571 - `"claude-opus-4-5"`
2572
2573 Powerful intelligence for long-running agents and coding
2574
2575 - `"claude-opus-4-5-20251101"`
2576
2577 Powerful intelligence for long-running agents and coding
2578
2579 - `"claude-sonnet-4-5"`
2580
2581 High-performance model for agents and coding
2582
2583 - `"claude-sonnet-4-5-20250929"`
2584
2585 High-performance model for agents and coding
26202586
26212587 - `string`
26222588
from line 3164
31983164
31993165 - `string`
32003166
3201 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
3202
3203 - `"message-batches-2024-09-24"`
3204
3205 - `"prompt-caching-2024-07-31"`
3206
3207 - `"computer-use-2024-10-22"`
3208
3209 - `"computer-use-2025-01-24"`
3210
3211 - `"pdfs-2024-09-25"`
3212
3213 - `"token-counting-2024-11-01"`
3214
3215 - `"token-efficient-tools-2025-02-19"`
3216
3217 - `"output-128k-2025-02-19"`
3218
3219 - `"files-api-2025-04-14"`
3220
3221 - `"mcp-client-2025-04-04"`
3222
3223 - `"mcp-client-2025-11-20"`
3224
3225 - `"dev-full-thinking-2025-05-14"`
3226
3227 - `"interleaved-thinking-2025-05-14"`
3228
3229 - `"code-execution-2025-05-22"`
3230
3231 - `"extended-cache-ttl-2025-04-11"`
3232
3233 - `"context-1m-2025-08-07"`
3234
3235 - `"context-management-2025-06-27"`
3236
3237 - `"model-context-window-exceeded-2025-08-26"`
3238
3239 - `"skills-2025-10-02"`
3240
3241 - `"fast-mode-2026-02-01"`
3242
3243 - `"output-300k-2026-03-24"`
3244
3245 - `"user-profiles-2026-03-24"`
3246
3247 - `"user-profiles-2026-08-18"`
3248
3249 - `"user-profiles-2026-09-04"`
3250
3251 - `"advisor-tool-2026-03-01"`
3252
3253 - `"managed-agents-2026-04-01"`
3254
3255 - `"cache-diagnosis-2026-04-07"`
3256
3257 - `"dreaming-2026-04-21"`
3258
3259 - `"thinking-token-count-2026-05-13"`
3260
3261 - `"server-side-fallback-2026-06-01"`
3262
3263 - `"server-side-fallback-2026-07-01"`
3264
3265 - `"fallback-credit-2026-06-01"`
3266
3267 - `"fallback-credit-2026-07-01"`
3268
3269 - `"agent-memory-2026-07-22"`
3270
3271 - `"mid-conversation-tool-changes-2026-07-01"`
3272
3273 - `"compact-2026-01-12"`
3274
3275 - `"computer-use-2025-11-24"`
3276
3277 - `"mcp-tunnels-2026-06-22"`
3278
3279 - `"structured-outputs-2025-11-13"`
3280
3281 - `"task-budgets-2026-03-13"`
3282
3283 - `"thinking-display-updates-2026-08-18"`
3284
3285 - `"ce-user-management-2026-07-13"`
3286
3287 - `"mid-conversation-output-config-2026-07-01"`
3288
3289 - `"thinking-binding-controls-2026-08-01"`
3290
3291 - `"mid-conversation-system-clear-at-2026-08-21"`
3292
3293 - `"compact-2026-09-04"`
3294
3295 - `"inline-tools-2026-09-15"`
3296
3297 - `"mcp-client-2026-09-15"`
3167 - `"message-batches-2024-09-24"`
3168
3169 - `"prompt-caching-2024-07-31"`
3170
3171 - `"computer-use-2024-10-22"`
3172
3173 - `"computer-use-2025-01-24"`
3174
3175 - `"pdfs-2024-09-25"`
3176
3177 - `"token-counting-2024-11-01"`
3178
3179 - `"token-efficient-tools-2025-02-19"`
3180
3181 - `"output-128k-2025-02-19"`
3182
3183 - `"files-api-2025-04-14"`
3184
3185 - `"mcp-client-2025-04-04"`
3186
3187 - `"mcp-client-2025-11-20"`
3188
3189 - `"dev-full-thinking-2025-05-14"`
3190
3191 - `"interleaved-thinking-2025-05-14"`
3192
3193 - `"code-execution-2025-05-22"`
3194
3195 - `"extended-cache-ttl-2025-04-11"`
3196
3197 - `"context-1m-2025-08-07"`
3198
3199 - `"context-management-2025-06-27"`
3200
3201 - `"model-context-window-exceeded-2025-08-26"`
3202
3203 - `"skills-2025-10-02"`
3204
3205 - `"fast-mode-2026-02-01"`
3206
3207 - `"output-300k-2026-03-24"`
3208
3209 - `"user-profiles-2026-03-24"`
3210
3211 - `"user-profiles-2026-08-18"`
3212
3213 - `"user-profiles-2026-09-04"`
3214
3215 - `"advisor-tool-2026-03-01"`
3216
3217 - `"managed-agents-2026-04-01"`
3218
3219 - `"cache-diagnosis-2026-04-07"`
3220
3221 - `"dreaming-2026-04-21"`
3222
3223 - `"thinking-token-count-2026-05-13"`
3224
3225 - `"server-side-fallback-2026-06-01"`
3226
3227 - `"server-side-fallback-2026-07-01"`
3228
3229 - `"fallback-credit-2026-06-01"`
3230
3231 - `"fallback-credit-2026-07-01"`
3232
3233 - `"agent-memory-2026-07-22"`
3234
3235 - `"mid-conversation-tool-changes-2026-07-01"`
3236
3237 - `"compact-2026-01-12"`
3238
3239 - `"computer-use-2025-11-24"`
3240
3241 - `"mcp-tunnels-2026-06-22"`
3242
3243 - `"structured-outputs-2025-11-13"`
3244
3245 - `"task-budgets-2026-03-13"`
3246
3247 - `"thinking-display-updates-2026-08-18"`
3248
3249 - `"ce-user-management-2026-07-13"`
3250
3251 - `"mid-conversation-output-config-2026-07-01"`
3252
3253 - `"thinking-binding-controls-2026-08-01"`
3254
3255 - `"mid-conversation-system-clear-at-2026-08-21"`
3256
3257 - `"compact-2026-09-04"`
3258
3259 - `"inline-tools-2026-09-15"`
3260
3261 - `"mcp-client-2026-09-15"`
32983262
32993263- `"anthropic-workspace-id": optional string`
33003264
from line 3306
33423306
33433307 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
33443308
3345 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
3309 - `"claude-sonnet-5-5"`
3310
3311 Efficient model for coding and agents
3312
3313 - `"claude-opus-5-5"`
3314
3315 Powerful intelligence for coding, knowledge work, and long-running agents
3316
3317 - `"claude-fable-5-1"`
3318
3319 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
3320
3321 - `"claude-sonnet-5"`
3322
3323 Efficient model for coding and agents
3324
3325 - `"claude-fable-5"`
3326
3327 Next generation of intelligence for the hardest knowledge work and coding problems
3328
3329 - `"claude-opus-5"`
3330
3331 Powerful intelligence for long-running agents and coding
3332
3333 - `"claude-opus-4-8"`
3334
3335 Powerful intelligence for long-running agents and coding
3336
3337 - `"claude-opus-4-7"`
3338
3339 Powerful intelligence for long-running agents and coding
3340
3341 - `"claude-opus-4-6"`
3342
3343 Powerful intelligence for long-running agents and coding
3344
3345 - `"claude-sonnet-4-6"`
3346
3347 Best combination of speed and intelligence
3348
3349 - `"claude-haiku-4-5"`
3350
3351 Fastest model with near-frontier intelligence
3352
3353 - `"claude-haiku-4-5-20251001"`
3354
3355 Fastest model with near-frontier intelligence
3356
3357 - `"claude-opus-4-5"`
3358
3359 Powerful intelligence for long-running agents and coding
3360
3361 - `"claude-opus-4-5-20251101"`
3362
3363 Powerful intelligence for long-running agents and coding
3364
3365 - `"claude-sonnet-4-5"`
3366
3367 High-performance model for agents and coding
3368
3369 - `"claude-sonnet-4-5-20250929"`
3370
3371 High-performance model for agents and coding
3372
3373 - `string`
3374
3375 - `BetaManagedAgentsModelConfigParams object`
3376
3377 An object that defines additional configuration control over model use
3378
3379 - `id: BetaManagedAgentsModel`
33463380
33473381 The model that will power your agent.
33483382
33493383 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
33503384
3385 - `effort: optional "low" or "medium" or "high" or 2 more or BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or 3 more or null`
3386
3387 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
3388
3389 - `"low"`
3390
3391 Low effort. Favors latency over reasoning depth.
3392
3393 - `"medium"`
3394
3395 Medium effort. Balances latency and reasoning depth.
3396
3397 - `"high"`
3398
3399 High effort. Favors reasoning depth.
3400
3401 - `"xhigh"`
3402
3403 Extra-high effort. Not all models accept this level.
3404
3405 - `"max"`
3406
3407 Maximum effort. Favors reasoning depth over latency.
3408
3409 - `BetaManagedAgentsEffortLow object`
3410
3411 Low effort. Favors latency over reasoning depth.
3412
3413 - `type: "low"`
3414
3415 - `BetaManagedAgentsEffortMedium object`
3416
3417 Medium effort. Balances latency and reasoning depth.
3418
3419 - `type: "medium"`
3420
3421 - `BetaManagedAgentsEffortHigh object`
3422
3423 High effort. Favors reasoning depth.
3424
3425 - `type: "high"`
3426
3427 - `BetaManagedAgentsEffortXhigh object`
3428
3429 Extra-high effort. Not all models accept this level.
3430
3431 - `type: "xhigh"`
3432
3433 - `BetaManagedAgentsEffortMax object`
3434
3435 Maximum effort. Favors reasoning depth over latency.
3436
3437 - `type: "max"`
3438
3439 - `inference_geo: optional string or null`
3440
3441 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo. On update, `model` is whole-object replacement — omitting inference_geo clears it.
3442
3443 - `speed: optional "standard" or "fast" or null`
3444
3445 Inference speed mode. Defaults to `standard`.
3446
3447 - `"standard"`
3448
3449 - `"fast"`
3450
3451- `multiagent: optional BetaManagedAgentsMultiagentParams or null`
3452
3453 Multiagent orchestration configuration. Full replacement. Omit to preserve; send null to clear.
3454
3455 - `type: "coordinator"`
3456
3457 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
3458
3459 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
3460
3461 - `string`
3462
3463 - `BetaManagedAgentsAgentParams object`
3464
3465 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
3466
3467 - `type: "agent"`
3468
3469 - `id: string`
3470
3471 The `agent` ID.
3472
3473 minLength: 1, maxLength: 128
3474
3475 - `version: optional number`
3476
3477 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
3478
3479 format: int32
3480
3481 - `BetaManagedAgentsMultiagentSelfParams object`
3482
3483 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
3484
3485 - `type: "self"`
3486
3487 - `BetaManagedAgentsAdvisorParams object`
3488
3489 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
3490
3491 - `type: "advisor"`
3492
3493 - `model: string`
3494
3495 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
3496
3497 minLength: 1, maxLength: 256
3498
3499- `name: optional string`
3500
3501 Human-readable name. Must be non-empty. Omit to preserve. Cannot be cleared.
3502
3503 maxLength: 256
3504
3505- `skills: optional array of BetaManagedAgentsSkillParams or null`
3506
3507 Skills. Full replacement. Omit to preserve; send empty array or null to clear.
3508
3509 - `BetaManagedAgentsAnthropicSkillParams object`
3510
3511 An Anthropic-managed skill.
3512
3513 - `type: "anthropic"`
3514
3515 - `skill_id: string`
3516
3517 Identifier of the Anthropic skill (e.g., "xlsx").
3518
3519 minLength: 1, maxLength: 64
3520
3521 - `version: optional string or null`
3522
3523 Version to pin. Defaults to latest if omitted.
3524
3525 minLength: 1, maxLength: 64
3526
3527 - `BetaManagedAgentsCustomSkillParams object`
3528
3529 A user-created custom skill.
3530
3531 - `type: "custom"`
3532
3533 - `skill_id: string`
3534
3535 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
3536
3537 minLength: 1, maxLength: 64
3538
3539 - `version: optional string or null`
3540
3541 Version to pin. Defaults to latest if omitted.
3542
3543 minLength: 1, maxLength: 64
3544
3545- `system: optional string or null`
3546
3547 System prompt. Omit to preserve; send empty string or null to clear.
3548
3549 maxLength: 100000
3550
3551- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams or null`
3552
3553 Tool configurations available to the agent. Full replacement. Omit to preserve; send empty array or null to clear. Maximum of 128 tools across all toolsets allowed.
3554
3555 - `BetaManagedAgentsAgentToolset20260401Params object`
3556
3557 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
3558
3559 - `type: "agent_toolset_20260401"`
3560
3561 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
3562
3563 Per-tool configuration overrides.
3564
3565 - `BetaManagedAgentsBashToolConfigParams object`
3566
3567 Configuration override for the bash tool.
3568
3569 - `type: optional "bash"`
3570
3571 - `name: "bash"`
3572
3573 Must be "bash".
3574
3575 - `enabled: optional boolean or null`
3576
3577 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3578
3579 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3580
3581 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3582
3583 - `BetaManagedAgentsAlwaysAllowPolicy object`
3584
3585 Tool calls are automatically approved without user confirmation.
3586
3587 - `type: "always_allow"`
3588
3589 - `BetaManagedAgentsAlwaysAskPolicy object`
3590
3591 Tool calls require user confirmation before execution.
3592
3593 - `type: "always_ask"`
3594
3595 - `BetaManagedAgentsAutoPolicy object`
3596
3597 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3598
3599 - `type: "auto"`
3600
3601 - `BetaManagedAgentsEditToolConfigParams object`
3602
3603 Configuration override for the edit tool.
3604
3605 - `type: optional "edit"`
3606
3607 - `name: "edit"`
3608
3609 Must be "edit".
3610
3611 - `enabled: optional boolean or null`
3612
3613 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3614
3615 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3616
3617 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3618
3619 - `BetaManagedAgentsAlwaysAllowPolicy object`
3620
3621 Tool calls are automatically approved without user confirmation.
3622
3623 - `BetaManagedAgentsAlwaysAskPolicy object`
3624
3625 Tool calls require user confirmation before execution.
3626
3627 - `BetaManagedAgentsAutoPolicy object`
3628
3629 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3630
3631 - `BetaManagedAgentsReadToolConfigParams object`
3632
3633 Configuration override for the read tool.
3634
3635 - `type: optional "read"`
3636
3637 - `name: "read"`
3638
3639 Must be "read".
3640
3641 - `enabled: optional boolean or null`
3642
3643 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3644
3645 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3646
3647 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3648
3649 - `BetaManagedAgentsAlwaysAllowPolicy object`
3650
3651 Tool calls are automatically approved without user confirmation.
3652
3653 - `BetaManagedAgentsAlwaysAskPolicy object`
3654
3655 Tool calls require user confirmation before execution.
3656
3657 - `BetaManagedAgentsAutoPolicy object`
3658
3659 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3660
3661 - `BetaManagedAgentsWriteToolConfigParams object`
3662
3663 Configuration override for the write tool.
3664
3665 - `type: optional "write"`
3666
3667 - `name: "write"`
3668
3669 Must be "write".
3670
3671 - `enabled: optional boolean or null`
3672
3673 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3674
3675 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3676
3677 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3678
3679 - `BetaManagedAgentsAlwaysAllowPolicy object`
3680
3681 Tool calls are automatically approved without user confirmation.
3682
3683 - `BetaManagedAgentsAlwaysAskPolicy object`
3684
3685 Tool calls require user confirmation before execution.
3686
3687 - `BetaManagedAgentsAutoPolicy object`
3688
3689 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3690
3691 - `BetaManagedAgentsGlobToolConfigParams object`
3692
3693 Configuration override for the glob tool.
3694
3695 - `type: optional "glob"`
3696
3697 - `name: "glob"`
3698
3699 Must be "glob".
3700
3701 - `enabled: optional boolean or null`
3702
3703 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3704
3705 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3706
3707 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3708
3709 - `BetaManagedAgentsAlwaysAllowPolicy object`
3710
3711 Tool calls are automatically approved without user confirmation.
3712
3713 - `BetaManagedAgentsAlwaysAskPolicy object`
3714
3715 Tool calls require user confirmation before execution.
3716
3717 - `BetaManagedAgentsAutoPolicy object`
3718
3719 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3720
3721 - `BetaManagedAgentsGrepToolConfigParams object`
3722
3723 Configuration override for the grep tool.
3724
3725 - `type: optional "grep"`
3726
3727 - `name: "grep"`
3728
3729 Must be "grep".
3730
3731 - `enabled: optional boolean or null`
3732
3733 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3734
3735 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3736
3737 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3738
3739 - `BetaManagedAgentsAlwaysAllowPolicy object`
3740
3741 Tool calls are automatically approved without user confirmation.
3742
3743 - `BetaManagedAgentsAlwaysAskPolicy object`
3744
3745 Tool calls require user confirmation before execution.
3746
3747 - `BetaManagedAgentsAutoPolicy object`
3748
3749 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3750
3751 - `BetaManagedAgentsWebFetchToolConfigParams object`
3752
3753 Configuration override for the web_fetch tool.
3754
3755 - `type: optional "web_fetch"`
3756
3757 - `name: "web_fetch"`
3758
3759 Must be "web_fetch".
3760
3761 - `allowed_domains: optional array of string`
3762
3763 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
3764
3765 - `blocked_domains: optional array of string`
3766
3767 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
3768
3769 - `enabled: optional boolean or null`
3770
3771 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3772
3773 - `max_content_tokens: optional number or null`
3774
3775 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
3776
3777 format: int32
3778
3779 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3780
3781 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3782
3783 - `BetaManagedAgentsAlwaysAllowPolicy object`
3784
3785 Tool calls are automatically approved without user confirmation.
3786
3787 - `BetaManagedAgentsAlwaysAskPolicy object`
3788
3789 Tool calls require user confirmation before execution.
3790
3791 - `BetaManagedAgentsAutoPolicy object`
3792
3793 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3794
3795 - `BetaManagedAgentsWebSearchToolConfigParams object`
3796
3797 Configuration override for the web_search tool.
3798
3799 - `type: optional "web_search"`
3800
3801 - `name: "web_search"`
3802
3803 Must be "web_search".
3804
3805 - `allowed_domains: optional array of string`
3806
3807 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
3808
3809 - `blocked_domains: optional array of string`
3810
3811 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
3812
3813 - `enabled: optional boolean or null`
3814
3815 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3816
3817 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3818
3819 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3820
3821 - `BetaManagedAgentsAlwaysAllowPolicy object`
3822
3823 Tool calls are automatically approved without user confirmation.
3824
3825 - `BetaManagedAgentsAlwaysAskPolicy object`
3826
3827 Tool calls require user confirmation before execution.
3828
3829 - `BetaManagedAgentsAutoPolicy object`
3830
3831 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3832
3833 - `user_location: optional BetaManagedAgentsUserLocation or null`
3834
3835 Approximate user location for search result localization.
3836
3837 - `type: "approximate"`
3838
3839 Location precision. Only "approximate" is supported.
3840
3841 - `city: optional string or null`
3842
3843 City name.
3844
3845 minLength: 1, maxLength: 255
3846
3847 - `country: optional string or null`
3848
3849 Two-letter ISO 3166-1 country code, uppercase.
3850
3851 - `region: optional string or null`
3852
3853 Region or state name.
3854
3855 minLength: 1, maxLength: 255
3856
3857 - `timezone: optional string or null`
3858
3859 IANA timezone identifier, e.g. "America/Los_Angeles".
3860
3861 minLength: 1, maxLength: 255
3862
3863 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
3864
3865 Default configuration applied to all tools in this set.
3866
3867 - `enabled: optional boolean or null`
3868
3869 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
3870
3871 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3872
3873 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
3874
3875 - `BetaManagedAgentsAlwaysAllowPolicy object`
3876
3877 Tool calls are automatically approved without user confirmation.
3878
3879 - `BetaManagedAgentsAlwaysAskPolicy object`
3880
3881 Tool calls require user confirmation before execution.
3882
3883 - `BetaManagedAgentsAutoPolicy object`
3884
3885 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3886
3887 - `BetaManagedAgentsMCPToolsetParams object`
3888
3889 Configuration for tools from an MCP server defined in `mcp_servers`.
3890
3891 - `type: "mcp_toolset"`
3892
3893 - `mcp_server_name: string`
3894
3895 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
3896
3897 minLength: 1, maxLength: 255
3898
3899 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
3900
3901 Per-tool configuration overrides.
3902
3903 - `name: string`
3904
3905 Name of the MCP tool to configure. 1-128 characters.
3906
3907 minLength: 1, maxLength: 128
3908
3909 - `enabled: optional boolean or null`
3910
3911 Whether this tool is enabled. Overrides the `default_config` setting.
3912
3913 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3914
3915 Permission policy for this tool. Overrides the `default_config` setting.
3916
3917 - `BetaManagedAgentsAlwaysAllowPolicy object`
3918
3919 Tool calls are automatically approved without user confirmation.
3920
3921 - `BetaManagedAgentsAlwaysAskPolicy object`
3922
3923 Tool calls require user confirmation before execution.
3924
3925 - `BetaManagedAgentsAutoPolicy object`
3926
3927 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3928
3929 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
3930
3931 Default configuration for all tools from this server.
3932
3933 - `enabled: optional boolean or null`
3934
3935 Whether tools are enabled by default. Defaults to true if not specified.
3936
3937 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3938
3939 Default permission policy for tools from this server.
3940
3941 - `BetaManagedAgentsAlwaysAllowPolicy object`
3942
3943 Tool calls are automatically approved without user confirmation.
3944
3945 - `BetaManagedAgentsAlwaysAskPolicy object`
3946
3947 Tool calls require user confirmation before execution.
3948
3949 - `BetaManagedAgentsAutoPolicy object`
3950
3951 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3952
3953 - `BetaManagedAgentsCustomToolParams object`
3954
3955 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
3956
3957 - `type: "custom"`
3958
3959 - `description: string`
3960
3961 Description of what the tool does, shown to the agent to help it decide when to use the tool.
3962
3963 minLength: 1
3964
3965 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
3966
3967 JSON Schema defining the expected input parameters for the tool.
3968
3969 - `type: "object"`
3970
3971 - `properties: optional map[unknown] or null`
3972
3973 - `required: optional array of string or null`
3974
3975 - `name: string`
3976
3977 Unique name for the tool. 1-128 characters; letters, digits, underscores, and hyphens.
3978
3979 minLength: 1, maxLength: 128
3980
3981- `version: optional number`
3982
3983 The agent's current version, used to prevent concurrent overwrites. Obtain this value from a create or retrieve response. Must be at least 1 if specified. When supplied, the request fails if it does not match the server's current version; omit to apply the update unconditionally.
3984
3985 format: int32
3986
3987### Returns
3988
3989- `BetaManagedAgentsAgent object`
3990
3991 A Managed Agents `agent`.
3992
3993 - `type: "agent"`
3994
3995 - `id: string`
3996
3997 - `archived_at: string or null`
3998
3999 When the agent was archived. Null if not archived.
4000
4001 format: date-time
4002
4003 - `created_at: string`
4004
4005 A timestamp in RFC 3339 format
4006
4007 format: date-time
4008
4009 - `description: string or null`
4010
4011 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
4012
4013 - `type: "url"`
4014
4015 - `name: string`
4016
4017 - `url: string`
4018
4019 - `metadata: map[string]`
4020
4021 - `model: BetaManagedAgentsModelConfig`
4022
4023 Model identifier and configuration.
4024
4025 - `id: BetaManagedAgentsModel`
4026
4027 The model that will power your agent.
4028
4029 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
4030
33514031 - `"claude-sonnet-5-5"`
33524032
33534033 Efficient model for coding and agents
from line 4091
34114091 - `"claude-sonnet-4-5-20250929"`
34124092
34134093 High-performance model for agents and coding
3414
3415 - `string`
3416
3417 - `BetaManagedAgentsModelConfigParams object`
3418
3419 An object that defines additional configuration control over model use
3420
3421 - `id: BetaManagedAgentsModel`
3422
3423 The model that will power your agent.
3424
3425 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
3426
3427 - `effort: optional "low" or "medium" or "high" or 2 more or BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or 3 more or null`
3428
3429 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
3430
3431 - `BetaManagedAgentsEffortLevel = "low" or "medium" or "high" or 2 more`
3432
3433 How hard Claude works on each turn. Higher levels favor reasoning depth over latency. Not all models accept every level; invalid combinations are rejected at create time.
3434
3435 - `"low"`
3436
3437 Low effort. Favors latency over reasoning depth.
3438
3439 - `"medium"`
3440
3441 Medium effort. Balances latency and reasoning depth.
3442
3443 - `"high"`
3444
3445 High effort. Favors reasoning depth.
3446
3447 - `"xhigh"`
3448
3449 Extra-high effort. Not all models accept this level.
3450
3451 - `"max"`
3452
3453 Maximum effort. Favors reasoning depth over latency.
3454
3455 - `BetaManagedAgentsEffortLow object`
3456
3457 Low effort. Favors latency over reasoning depth.
3458
3459 - `type: "low"`
3460
3461 - `BetaManagedAgentsEffortMedium object`
3462
3463 Medium effort. Balances latency and reasoning depth.
3464
3465 - `type: "medium"`
3466
3467 - `BetaManagedAgentsEffortHigh object`
3468
3469 High effort. Favors reasoning depth.
3470
3471 - `type: "high"`
3472
3473 - `BetaManagedAgentsEffortXhigh object`
3474
3475 Extra-high effort. Not all models accept this level.
3476
3477 - `type: "xhigh"`
3478
3479 - `BetaManagedAgentsEffortMax object`
3480
3481 Maximum effort. Favors reasoning depth over latency.
3482
3483 - `type: "max"`
3484
3485 - `inference_geo: optional string or null`
3486
3487 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo. On update, `model` is whole-object replacement — omitting inference_geo clears it.
3488
3489 - `speed: optional "standard" or "fast" or null`
3490
3491 Inference speed mode. Defaults to `standard`.
3492
3493 - `"standard"`
3494
3495 - `"fast"`
3496
3497- `multiagent: optional BetaManagedAgentsMultiagentParams or null`
3498
3499 Multiagent orchestration configuration. Full replacement. Omit to preserve; send null to clear.
3500
3501 - `type: "coordinator"`
3502
3503 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
3504
3505 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
3506
3507 - `string`
3508
3509 - `BetaManagedAgentsAgentParams object`
3510
3511 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
3512
3513 - `type: "agent"`
3514
3515 - `id: string`
3516
3517 The `agent` ID.
3518
3519 minLength: 1, maxLength: 128
3520
3521 - `version: optional number`
3522
3523 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
3524
3525 format: int32
3526
3527 - `BetaManagedAgentsMultiagentSelfParams object`
3528
3529 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
3530
3531 - `type: "self"`
3532
3533 - `BetaManagedAgentsAdvisorParams object`
3534
3535 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
3536
3537 - `type: "advisor"`
3538
3539 - `model: string`
3540
3541 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
3542
3543 minLength: 1, maxLength: 256
3544
3545- `name: optional string`
3546
3547 Human-readable name. Must be non-empty. Omit to preserve. Cannot be cleared.
3548
3549 maxLength: 256
3550
3551- `skills: optional array of BetaManagedAgentsSkillParams or null`
3552
3553 Skills. Full replacement. Omit to preserve; send empty array or null to clear.
3554
3555 - `BetaManagedAgentsAnthropicSkillParams object`
3556
3557 An Anthropic-managed skill.
3558
3559 - `type: "anthropic"`
3560
3561 - `skill_id: string`
3562
3563 Identifier of the Anthropic skill (e.g., "xlsx").
3564
3565 minLength: 1, maxLength: 64
3566
3567 - `version: optional string or null`
3568
3569 Version to pin. Defaults to latest if omitted.
3570
3571 minLength: 1, maxLength: 64
3572
3573 - `BetaManagedAgentsCustomSkillParams object`
3574
3575 A user-created custom skill.
3576
3577 - `type: "custom"`
3578
3579 - `skill_id: string`
3580
3581 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
3582
3583 minLength: 1, maxLength: 64
3584
3585 - `version: optional string or null`
3586
3587 Version to pin. Defaults to latest if omitted.
3588
3589 minLength: 1, maxLength: 64
3590
3591- `system: optional string or null`
3592
3593 System prompt. Omit to preserve; send empty string or null to clear.
3594
3595 maxLength: 100000
3596
3597- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams or null`
3598
3599 Tool configurations available to the agent. Full replacement. Omit to preserve; send empty array or null to clear. Maximum of 128 tools across all toolsets allowed.
3600
3601 - `BetaManagedAgentsAgentToolset20260401Params object`
3602
3603 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
3604
3605 - `type: "agent_toolset_20260401"`
3606
3607 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
3608
3609 Per-tool configuration overrides.
3610
3611 - `BetaManagedAgentsBashToolConfigParams object`
3612
3613 Configuration override for the bash tool.
3614
3615 - `type: optional "bash"`
3616
3617 - `name: "bash"`
3618
3619 Must be "bash".
3620
3621 - `enabled: optional boolean or null`
3622
3623 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3624
3625 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3626
3627 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3628
3629 - `BetaManagedAgentsAlwaysAllowPolicy object`
3630
3631 Tool calls are automatically approved without user confirmation.
3632
3633 - `type: "always_allow"`
3634
3635 - `BetaManagedAgentsAlwaysAskPolicy object`
3636
3637 Tool calls require user confirmation before execution.
3638
3639 - `type: "always_ask"`
3640
3641 - `BetaManagedAgentsAutoPolicy object`
3642
3643 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3644
3645 - `type: "auto"`
3646
3647 - `BetaManagedAgentsEditToolConfigParams object`
3648
3649 Configuration override for the edit tool.
3650
3651 - `type: optional "edit"`
3652
3653 - `name: "edit"`
3654
3655 Must be "edit".
3656
3657 - `enabled: optional boolean or null`
3658
3659 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3660
3661 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3662
3663 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3664
3665 - `BetaManagedAgentsAlwaysAllowPolicy object`
3666
3667 Tool calls are automatically approved without user confirmation.
3668
3669 - `BetaManagedAgentsAlwaysAskPolicy object`
3670
3671 Tool calls require user confirmation before execution.
3672
3673 - `BetaManagedAgentsAutoPolicy object`
3674
3675 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3676
3677 - `BetaManagedAgentsReadToolConfigParams object`
3678
3679 Configuration override for the read tool.
3680
3681 - `type: optional "read"`
3682
3683 - `name: "read"`
3684
3685 Must be "read".
3686
3687 - `enabled: optional boolean or null`
3688
3689 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3690
3691 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3692
3693 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3694
3695 - `BetaManagedAgentsAlwaysAllowPolicy object`
3696
3697 Tool calls are automatically approved without user confirmation.
3698
3699 - `BetaManagedAgentsAlwaysAskPolicy object`
3700
3701 Tool calls require user confirmation before execution.
3702
3703 - `BetaManagedAgentsAutoPolicy object`
3704
3705 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3706
3707 - `BetaManagedAgentsWriteToolConfigParams object`
3708
3709 Configuration override for the write tool.
3710
3711 - `type: optional "write"`
3712
3713 - `name: "write"`
3714
3715 Must be "write".
3716
3717 - `enabled: optional boolean or null`
3718
3719 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3720
3721 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3722
3723 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3724
3725 - `BetaManagedAgentsAlwaysAllowPolicy object`
3726
3727 Tool calls are automatically approved without user confirmation.
3728
3729 - `BetaManagedAgentsAlwaysAskPolicy object`
3730
3731 Tool calls require user confirmation before execution.
3732
3733 - `BetaManagedAgentsAutoPolicy object`
3734
3735 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3736
3737 - `BetaManagedAgentsGlobToolConfigParams object`
3738
3739 Configuration override for the glob tool.
3740
3741 - `type: optional "glob"`
3742
3743 - `name: "glob"`
3744
3745 Must be "glob".
3746
3747 - `enabled: optional boolean or null`
3748
3749 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3750
3751 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3752
3753 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3754
3755 - `BetaManagedAgentsAlwaysAllowPolicy object`
3756
3757 Tool calls are automatically approved without user confirmation.
3758
3759 - `BetaManagedAgentsAlwaysAskPolicy object`
3760
3761 Tool calls require user confirmation before execution.
3762
3763 - `BetaManagedAgentsAutoPolicy object`
3764
3765 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3766
3767 - `BetaManagedAgentsGrepToolConfigParams object`
3768
3769 Configuration override for the grep tool.
3770
3771 - `type: optional "grep"`
3772
3773 - `name: "grep"`
3774
3775 Must be "grep".
3776
3777 - `enabled: optional boolean or null`
3778
3779 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3780
3781 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3782
3783 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3784
3785 - `BetaManagedAgentsAlwaysAllowPolicy object`
3786
3787 Tool calls are automatically approved without user confirmation.
3788
3789 - `BetaManagedAgentsAlwaysAskPolicy object`
3790
3791 Tool calls require user confirmation before execution.
3792
3793 - `BetaManagedAgentsAutoPolicy object`
3794
3795 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3796
3797 - `BetaManagedAgentsWebFetchToolConfigParams object`
3798
3799 Configuration override for the web_fetch tool.
3800
3801 - `type: optional "web_fetch"`
3802
3803 - `name: "web_fetch"`
3804
3805 Must be "web_fetch".
3806
3807 - `allowed_domains: optional array of string`
3808
3809 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
3810
3811 - `blocked_domains: optional array of string`
3812
3813 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
3814
3815 - `enabled: optional boolean or null`
3816
3817 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3818
3819 - `max_content_tokens: optional number or null`
3820
3821 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
3822
3823 format: int32
3824
3825 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3826
3827 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3828
3829 - `BetaManagedAgentsAlwaysAllowPolicy object`
3830
3831 Tool calls are automatically approved without user confirmation.
3832
3833 - `BetaManagedAgentsAlwaysAskPolicy object`
3834
3835 Tool calls require user confirmation before execution.
3836
3837 - `BetaManagedAgentsAutoPolicy object`
3838
3839 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3840
3841 - `BetaManagedAgentsWebSearchToolConfigParams object`
3842
3843 Configuration override for the web_search tool.
3844
3845 - `type: optional "web_search"`
3846
3847 - `name: "web_search"`
3848
3849 Must be "web_search".
3850
3851 - `allowed_domains: optional array of string`
3852
3853 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
3854
3855 - `blocked_domains: optional array of string`
3856
3857 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
3858
3859 - `enabled: optional boolean or null`
3860
3861 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3862
3863 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3864
3865 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3866
3867 - `BetaManagedAgentsAlwaysAllowPolicy object`
3868
3869 Tool calls are automatically approved without user confirmation.
3870
3871 - `BetaManagedAgentsAlwaysAskPolicy object`
3872
3873 Tool calls require user confirmation before execution.
3874
3875 - `BetaManagedAgentsAutoPolicy object`
3876
3877 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3878
3879 - `user_location: optional BetaManagedAgentsUserLocation or null`
3880
3881 Approximate user location for search result localization.
3882
3883 - `type: "approximate"`
3884
3885 Location precision. Only "approximate" is supported.
3886
3887 - `city: optional string or null`
3888
3889 City name.
3890
3891 minLength: 1, maxLength: 255
3892
3893 - `country: optional string or null`
3894
3895 Two-letter ISO 3166-1 country code, uppercase.
3896
3897 - `region: optional string or null`
3898
3899 Region or state name.
3900
3901 minLength: 1, maxLength: 255
3902
3903 - `timezone: optional string or null`
3904
3905 IANA timezone identifier, e.g. "America/Los_Angeles".
3906
3907 minLength: 1, maxLength: 255
3908
3909 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
3910
3911 Default configuration applied to all tools in this set.
3912
3913 - `enabled: optional boolean or null`
3914
3915 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
3916
3917 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3918
3919 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
3920
3921 - `BetaManagedAgentsAlwaysAllowPolicy object`
3922
3923 Tool calls are automatically approved without user confirmation.
3924
3925 - `BetaManagedAgentsAlwaysAskPolicy object`
3926
3927 Tool calls require user confirmation before execution.
3928
3929 - `BetaManagedAgentsAutoPolicy object`
3930
3931 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3932
3933 - `BetaManagedAgentsMCPToolsetParams object`
3934
3935 Configuration for tools from an MCP server defined in `mcp_servers`.
3936
3937 - `type: "mcp_toolset"`
3938
3939 - `mcp_server_name: string`
3940
3941 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
3942
3943 minLength: 1, maxLength: 255
3944
3945 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
3946
3947 Per-tool configuration overrides.
3948
3949 - `name: string`
3950
3951 Name of the MCP tool to configure. 1-128 characters.
3952
3953 minLength: 1, maxLength: 128
3954
3955 - `enabled: optional boolean or null`
3956
3957 Whether this tool is enabled. Overrides the `default_config` setting.
3958
3959 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3960
3961 Permission policy for this tool. Overrides the `default_config` setting.
3962
3963 - `BetaManagedAgentsAlwaysAllowPolicy object`
3964
3965 Tool calls are automatically approved without user confirmation.
3966
3967 - `BetaManagedAgentsAlwaysAskPolicy object`
3968
3969 Tool calls require user confirmation before execution.
3970
3971 - `BetaManagedAgentsAutoPolicy object`
3972
3973 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3974
3975 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
3976
3977 Default configuration for all tools from this server.
3978
3979 - `enabled: optional boolean or null`
3980
3981 Whether tools are enabled by default. Defaults to true if not specified.
3982
3983 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3984
3985 Default permission policy for tools from this server.
3986
3987 - `BetaManagedAgentsAlwaysAllowPolicy object`
3988
3989 Tool calls are automatically approved without user confirmation.
3990
3991 - `BetaManagedAgentsAlwaysAskPolicy object`
3992
3993 Tool calls require user confirmation before execution.
3994
3995 - `BetaManagedAgentsAutoPolicy object`
3996
3997 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3998
3999 - `BetaManagedAgentsCustomToolParams object`
4000
4001 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
4002
4003 - `type: "custom"`
4004
4005 - `description: string`
4006
4007 Description of what the tool does, shown to the agent to help it decide when to use the tool.
4008
4009 minLength: 1
4010
4011 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
4012
4013 JSON Schema defining the expected input parameters for the tool.
4014
4015 - `type: "object"`
4016
4017 - `properties: optional map[unknown] or null`
4018
4019 - `required: optional array of string or null`
4020
4021 - `name: string`
4022
4023 Unique name for the tool. 1-128 characters; letters, digits, underscores, and hyphens.
4024
4025 minLength: 1, maxLength: 128
4026
4027- `version: optional number`
4028
4029 The agent's current version, used to prevent concurrent overwrites. Obtain this value from a create or retrieve response. Must be at least 1 if specified. When supplied, the request fails if it does not match the server's current version; omit to apply the update unconditionally.
4030
4031 format: int32
4032
4033### Returns
4034
4035- `BetaManagedAgentsAgent object`
4036
4037 A Managed Agents `agent`.
4038
4039 - `type: "agent"`
4040
4041 - `id: string`
4042
4043 - `archived_at: string or null`
4044
4045 When the agent was archived. Null if not archived.
4046
4047 format: date-time
4048
4049 - `created_at: string`
4050
4051 A timestamp in RFC 3339 format
4052
4053 format: date-time
4054
4055 - `description: string or null`
4056
4057 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
4058
4059 - `type: "url"`
4060
4061 - `name: string`
4062
4063 - `url: string`
4064
4065 - `metadata: map[string]`
4066
4067 - `model: BetaManagedAgentsModelConfig`
4068
4069 Model identifier and configuration.
4070
4071 - `id: BetaManagedAgentsModel`
4072
4073 The model that will power your agent.
4074
4075 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
4076
4077 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
4078
4079 The model that will power your agent.
4080
4081 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
4082
4083 - `"claude-sonnet-5-5"`
4084
4085 Efficient model for coding and agents
4086
4087 - `"claude-opus-5-5"`
4088
4089 Powerful intelligence for coding, knowledge work, and long-running agents
4090
4091 - `"claude-fable-5-1"`
4092
4093 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
4094
4095 - `"claude-sonnet-5"`
4096
4097 Efficient model for coding and agents
4098
4099 - `"claude-fable-5"`
4100
4101 Next generation of intelligence for the hardest knowledge work and coding problems
4102
4103 - `"claude-opus-5"`
4104
4105 Powerful intelligence for long-running agents and coding
4106
4107 - `"claude-opus-4-8"`
4108
4109 Powerful intelligence for long-running agents and coding
4110
4111 - `"claude-opus-4-7"`
4112
4113 Powerful intelligence for long-running agents and coding
4114
4115 - `"claude-opus-4-6"`
4116
4117 Powerful intelligence for long-running agents and coding
4118
4119 - `"claude-sonnet-4-6"`
4120
4121 Best combination of speed and intelligence
4122
4123 - `"claude-haiku-4-5"`
4124
4125 Fastest model with near-frontier intelligence
4126
4127 - `"claude-haiku-4-5-20251001"`
4128
4129 Fastest model with near-frontier intelligence
4130
4131 - `"claude-opus-4-5"`
4132
4133 Powerful intelligence for long-running agents and coding
4134
4135 - `"claude-opus-4-5-20251101"`
4136
4137 Powerful intelligence for long-running agents and coding
4138
4139 - `"claude-sonnet-4-5"`
4140
4141 High-performance model for agents and coding
4142
4143 - `"claude-sonnet-4-5-20250929"`
4144
4145 High-performance model for agents and coding
41464094
41474095 - `string`
41484096
from line 4687
47394687
47404688 - `string`
47414689
4742 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
4743
4744 - `"message-batches-2024-09-24"`
4745
4746 - `"prompt-caching-2024-07-31"`
4747
4748 - `"computer-use-2024-10-22"`
4749
4750 - `"computer-use-2025-01-24"`
4751
4752 - `"pdfs-2024-09-25"`
4753
4754 - `"token-counting-2024-11-01"`
4755
4756 - `"token-efficient-tools-2025-02-19"`
4757
4758 - `"output-128k-2025-02-19"`
4759
4760 - `"files-api-2025-04-14"`
4761
4762 - `"mcp-client-2025-04-04"`
4763
4764 - `"mcp-client-2025-11-20"`
4765
4766 - `"dev-full-thinking-2025-05-14"`
4767
4768 - `"interleaved-thinking-2025-05-14"`
4769
4770 - `"code-execution-2025-05-22"`
4771
4772 - `"extended-cache-ttl-2025-04-11"`
4773
4774 - `"context-1m-2025-08-07"`
4775
4776 - `"context-management-2025-06-27"`
4777
4778 - `"model-context-window-exceeded-2025-08-26"`
4779
4780 - `"skills-2025-10-02"`
4781
4782 - `"fast-mode-2026-02-01"`
4783
4784 - `"output-300k-2026-03-24"`
4785
4786 - `"user-profiles-2026-03-24"`
4787
4788 - `"user-profiles-2026-08-18"`
4789
4790 - `"user-profiles-2026-09-04"`
4791
4792 - `"advisor-tool-2026-03-01"`
4793
4794 - `"managed-agents-2026-04-01"`
4795
4796 - `"cache-diagnosis-2026-04-07"`
4797
4798 - `"dreaming-2026-04-21"`
4799
4800 - `"thinking-token-count-2026-05-13"`
4801
4802 - `"server-side-fallback-2026-06-01"`
4803
4804 - `"server-side-fallback-2026-07-01"`
4805
4806 - `"fallback-credit-2026-06-01"`
4807
4808 - `"fallback-credit-2026-07-01"`
4809
4810 - `"agent-memory-2026-07-22"`
4811
4812 - `"mid-conversation-tool-changes-2026-07-01"`
4813
4814 - `"compact-2026-01-12"`
4815
4816 - `"computer-use-2025-11-24"`
4817
4818 - `"mcp-tunnels-2026-06-22"`
4819
4820 - `"structured-outputs-2025-11-13"`
4821
4822 - `"task-budgets-2026-03-13"`
4823
4824 - `"thinking-display-updates-2026-08-18"`
4825
4826 - `"ce-user-management-2026-07-13"`
4827
4828 - `"mid-conversation-output-config-2026-07-01"`
4829
4830 - `"thinking-binding-controls-2026-08-01"`
4831
4832 - `"mid-conversation-system-clear-at-2026-08-21"`
4833
4834 - `"compact-2026-09-04"`
4835
4836 - `"inline-tools-2026-09-15"`
4837
4838 - `"mcp-client-2026-09-15"`
4690 - `"message-batches-2024-09-24"`
4691
4692 - `"prompt-caching-2024-07-31"`
4693
4694 - `"computer-use-2024-10-22"`
4695
4696 - `"computer-use-2025-01-24"`
4697
4698 - `"pdfs-2024-09-25"`
4699
4700 - `"token-counting-2024-11-01"`
4701
4702 - `"token-efficient-tools-2025-02-19"`
4703
4704 - `"output-128k-2025-02-19"`
4705
4706 - `"files-api-2025-04-14"`
4707
4708 - `"mcp-client-2025-04-04"`
4709
4710 - `"mcp-client-2025-11-20"`
4711
4712 - `"dev-full-thinking-2025-05-14"`
4713
4714 - `"interleaved-thinking-2025-05-14"`
4715
4716 - `"code-execution-2025-05-22"`
4717
4718 - `"extended-cache-ttl-2025-04-11"`
4719
4720 - `"context-1m-2025-08-07"`
4721
4722 - `"context-management-2025-06-27"`
4723
4724 - `"model-context-window-exceeded-2025-08-26"`
4725
4726 - `"skills-2025-10-02"`
4727
4728 - `"fast-mode-2026-02-01"`
4729
4730 - `"output-300k-2026-03-24"`
4731
4732 - `"user-profiles-2026-03-24"`
4733
4734 - `"user-profiles-2026-08-18"`
4735
4736 - `"user-profiles-2026-09-04"`
4737
4738 - `"advisor-tool-2026-03-01"`
4739
4740 - `"managed-agents-2026-04-01"`
4741
4742 - `"cache-diagnosis-2026-04-07"`
4743
4744 - `"dreaming-2026-04-21"`
4745
4746 - `"thinking-token-count-2026-05-13"`
4747
4748 - `"server-side-fallback-2026-06-01"`
4749
4750 - `"server-side-fallback-2026-07-01"`
4751
4752 - `"fallback-credit-2026-06-01"`
4753
4754 - `"fallback-credit-2026-07-01"`
4755
4756 - `"agent-memory-2026-07-22"`
4757
4758 - `"mid-conversation-tool-changes-2026-07-01"`
4759
4760 - `"compact-2026-01-12"`
4761
4762 - `"computer-use-2025-11-24"`
4763
4764 - `"mcp-tunnels-2026-06-22"`
4765
4766 - `"structured-outputs-2025-11-13"`
4767
4768 - `"task-budgets-2026-03-13"`
4769
4770 - `"thinking-display-updates-2026-08-18"`
4771
4772 - `"ce-user-management-2026-07-13"`
4773
4774 - `"mid-conversation-output-config-2026-07-01"`
4775
4776 - `"thinking-binding-controls-2026-08-01"`
4777
4778 - `"mid-conversation-system-clear-at-2026-08-21"`
4779
4780 - `"compact-2026-09-04"`
4781
4782 - `"inline-tools-2026-09-15"`
4783
4784 - `"mcp-client-2026-09-15"`
48394785
48404786- `"anthropic-workspace-id": optional string`
48414787
from line 4833
48874833
48884834 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
48894835
4890 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
4891
4892 The model that will power your agent.
4893
4894 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
4895
4896 - `"claude-sonnet-5-5"`
4897
4898 Efficient model for coding and agents
4899
4900 - `"claude-opus-5-5"`
4901
4902 Powerful intelligence for coding, knowledge work, and long-running agents
4903
4904 - `"claude-fable-5-1"`
4905
4906 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
4907
4908 - `"claude-sonnet-5"`
4909
4910 Efficient model for coding and agents
4911
4912 - `"claude-fable-5"`
4913
4914 Next generation of intelligence for the hardest knowledge work and coding problems
4915
4916 - `"claude-opus-5"`
4917
4918 Powerful intelligence for long-running agents and coding
4919
4920 - `"claude-opus-4-8"`
4921
4922 Powerful intelligence for long-running agents and coding
4923
4924 - `"claude-opus-4-7"`
4925
4926 Powerful intelligence for long-running agents and coding
4927
4928 - `"claude-opus-4-6"`
4929
4930 Powerful intelligence for long-running agents and coding
4931
4932 - `"claude-sonnet-4-6"`
4933
4934 Best combination of speed and intelligence
4935
4936 - `"claude-haiku-4-5"`
4937
4938 Fastest model with near-frontier intelligence
4939
4940 - `"claude-haiku-4-5-20251001"`
4941
4942 Fastest model with near-frontier intelligence
4943
4944 - `"claude-opus-4-5"`
4945
4946 Powerful intelligence for long-running agents and coding
4947
4948 - `"claude-opus-4-5-20251101"`
4949
4950 Powerful intelligence for long-running agents and coding
4951
4952 - `"claude-sonnet-4-5"`
4953
4954 High-performance model for agents and coding
4955
4956 - `"claude-sonnet-4-5-20250929"`
4957
4958 High-performance model for agents and coding
4836 - `"claude-sonnet-5-5"`
4837
4838 Efficient model for coding and agents
4839
4840 - `"claude-opus-5-5"`
4841
4842 Powerful intelligence for coding, knowledge work, and long-running agents
4843
4844 - `"claude-fable-5-1"`
4845
4846 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
4847
4848 - `"claude-sonnet-5"`
4849
4850 Efficient model for coding and agents
4851
4852 - `"claude-fable-5"`
4853
4854 Next generation of intelligence for the hardest knowledge work and coding problems
4855
4856 - `"claude-opus-5"`
4857
4858 Powerful intelligence for long-running agents and coding
4859
4860 - `"claude-opus-4-8"`
4861
4862 Powerful intelligence for long-running agents and coding
4863
4864 - `"claude-opus-4-7"`
4865
4866 Powerful intelligence for long-running agents and coding
4867
4868 - `"claude-opus-4-6"`
4869
4870 Powerful intelligence for long-running agents and coding
4871
4872 - `"claude-sonnet-4-6"`
4873
4874 Best combination of speed and intelligence
4875
4876 - `"claude-haiku-4-5"`
4877
4878 Fastest model with near-frontier intelligence
4879
4880 - `"claude-haiku-4-5-20251001"`
4881
4882 Fastest model with near-frontier intelligence
4883
4884 - `"claude-opus-4-5"`
4885
4886 Powerful intelligence for long-running agents and coding
4887
4888 - `"claude-opus-4-5-20251101"`
4889
4890 Powerful intelligence for long-running agents and coding
4891
4892 - `"claude-sonnet-4-5"`
4893
4894 High-performance model for agents and coding
4895
4896 - `"claude-sonnet-4-5-20250929"`
4897
4898 High-performance model for agents and coding
49594899
49604900 - `string`
49614901
from line 5516
55765516
55775517 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
55785518
5579 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
5580
5581 The model that will power your agent.
5582
5583 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
5584
5585 - `"claude-sonnet-5-5"`
5586
5587 Efficient model for coding and agents
5588
5589 - `"claude-opus-5-5"`
5590
5591 Powerful intelligence for coding, knowledge work, and long-running agents
5592
5593 - `"claude-fable-5-1"`
5594
5595 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
5596
5597 - `"claude-sonnet-5"`
5598
5599 Efficient model for coding and agents
5600
5601 - `"claude-fable-5"`
5602
5603 Next generation of intelligence for the hardest knowledge work and coding problems
5604
5605 - `"claude-opus-5"`
5606
5607 Powerful intelligence for long-running agents and coding
5608
5609 - `"claude-opus-4-8"`
5610
5611 Powerful intelligence for long-running agents and coding
5612
5613 - `"claude-opus-4-7"`
5614
5615 Powerful intelligence for long-running agents and coding
5616
5617 - `"claude-opus-4-6"`
5618
5619 Powerful intelligence for long-running agents and coding
5620
5621 - `"claude-sonnet-4-6"`
5622
5623 Best combination of speed and intelligence
5624
5625 - `"claude-haiku-4-5"`
5626
5627 Fastest model with near-frontier intelligence
5628
5629 - `"claude-haiku-4-5-20251001"`
5630
5631 Fastest model with near-frontier intelligence
5632
5633 - `"claude-opus-4-5"`
5634
5635 Powerful intelligence for long-running agents and coding
5636
5637 - `"claude-opus-4-5-20251101"`
5638
5639 Powerful intelligence for long-running agents and coding
5640
5641 - `"claude-sonnet-4-5"`
5642
5643 High-performance model for agents and coding
5644
5645 - `"claude-sonnet-4-5-20250929"`
5646
5647 High-performance model for agents and coding
5519 - `"claude-sonnet-5-5"`
5520
5521 Efficient model for coding and agents
5522
5523 - `"claude-opus-5-5"`
5524
5525 Powerful intelligence for coding, knowledge work, and long-running agents
5526
5527 - `"claude-fable-5-1"`
5528
5529 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
5530
5531 - `"claude-sonnet-5"`
5532
5533 Efficient model for coding and agents
5534
5535 - `"claude-fable-5"`
5536
5537 Next generation of intelligence for the hardest knowledge work and coding problems
5538
5539 - `"claude-opus-5"`
5540
5541 Powerful intelligence for long-running agents and coding
5542
5543 - `"claude-opus-4-8"`
5544
5545 Powerful intelligence for long-running agents and coding
5546
5547 - `"claude-opus-4-7"`
5548
5549 Powerful intelligence for long-running agents and coding
5550
5551 - `"claude-opus-4-6"`
5552
5553 Powerful intelligence for long-running agents and coding
5554
5555 - `"claude-sonnet-4-6"`
5556
5557 Best combination of speed and intelligence
5558
5559 - `"claude-haiku-4-5"`
5560
5561 Fastest model with near-frontier intelligence
5562
5563 - `"claude-haiku-4-5-20251001"`
5564
5565 Fastest model with near-frontier intelligence
5566
5567 - `"claude-opus-4-5"`
5568
5569 Powerful intelligence for long-running agents and coding
5570
5571 - `"claude-opus-4-5-20251101"`
5572
5573 Powerful intelligence for long-running agents and coding
5574
5575 - `"claude-sonnet-4-5"`
5576
5577 High-performance model for agents and coding
5578
5579 - `"claude-sonnet-4-5-20250929"`
5580
5581 High-performance model for agents and coding
56485582
56495583 - `string`
56505584
from line 8200
82668200
82678201 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
82688202
8269 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
8203 - `"claude-sonnet-5-5"`
8204
8205 Efficient model for coding and agents
8206
8207 - `"claude-opus-5-5"`
8208
8209 Powerful intelligence for coding, knowledge work, and long-running agents
8210
8211 - `"claude-fable-5-1"`
8212
8213 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
8214
8215 - `"claude-sonnet-5"`
8216
8217 Efficient model for coding and agents
8218
8219 - `"claude-fable-5"`
8220
8221 Next generation of intelligence for the hardest knowledge work and coding problems
8222
8223 - `"claude-opus-5"`
8224
8225 Powerful intelligence for long-running agents and coding
8226
8227 - `"claude-opus-4-8"`
8228
8229 Powerful intelligence for long-running agents and coding
8230
8231 - `"claude-opus-4-7"`
8232
8233 Powerful intelligence for long-running agents and coding
8234
8235 - `"claude-opus-4-6"`
8236
8237 Powerful intelligence for long-running agents and coding
8238
8239 - `"claude-sonnet-4-6"`
8240
8241 Best combination of speed and intelligence
8242
8243 - `"claude-haiku-4-5"`
8244
8245 Fastest model with near-frontier intelligence
8246
8247 - `"claude-haiku-4-5-20251001"`
8248
8249 Fastest model with near-frontier intelligence
8250
8251 - `"claude-opus-4-5"`
8252
8253 Powerful intelligence for long-running agents and coding
8254
8255 - `"claude-opus-4-5-20251101"`
8256
8257 Powerful intelligence for long-running agents and coding
8258
8259 - `"claude-sonnet-4-5"`
8260
8261 High-performance model for agents and coding
8262
8263 - `"claude-sonnet-4-5-20250929"`
8264
8265 High-performance model for agents and coding
8266
8267 - `string`
8268
8269### Beta Managed Agents Model Config
8270
8271- `BetaManagedAgentsModelConfig object`
8272
8273 Model identifier and configuration.
8274
8275 - `id: BetaManagedAgentsModel`
82708276
82718277 The model that will power your agent.
82728278
from line 8342
83368342
83378343 High-performance model for agents and coding
83388344
8345 - `string`
8346
8347 - `effort: optional BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or BetaManagedAgentsEffortHigh or 2 more`
8348
8349 How hard Claude works on each inference call. One of `low`, `medium`, `high`, `xhigh`, `max`. Always present; resolved to the per-model default at save time when not supplied.
8350
8351 - `BetaManagedAgentsEffortLow object`
8352
8353 Low effort. Favors latency over reasoning depth.
8354
8355 - `type: "low"`
8356
8357 - `BetaManagedAgentsEffortMedium object`
8358
8359 Medium effort. Balances latency and reasoning depth.
8360
8361 - `type: "medium"`
8362
8363 - `BetaManagedAgentsEffortHigh object`
8364
8365 High effort. Favors reasoning depth.
8366
8367 - `type: "high"`
8368
8369 - `BetaManagedAgentsEffortXhigh object`
8370
8371 Extra-high effort. Not all models accept this level.
8372
8373 - `type: "xhigh"`
8374
8375 - `BetaManagedAgentsEffortMax object`
8376
8377 Maximum effort. Favors reasoning depth over latency.
8378
8379 - `type: "max"`
8380
8381 - `inference_geo: optional string`
8382
8383 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo.
8384
8385 - `speed: optional "standard" or "fast"`
8386
8387 Inference speed mode. `fast` provides significantly faster output token generation at premium pricing. Defaults to `standard`. Not all models support `fast`; invalid combinations are rejected at create time.
8388
8389 - `"standard"`
8390
8391 - `"fast"`
8392
8393### Beta Managed Agents Model Config Params
8394
8395- `BetaManagedAgentsModelConfigParams object`
8396
8397 An object that defines additional configuration control over model use
8398
8399 - `id: BetaManagedAgentsModel`
8400
8401 The model that will power your agent.
8402
8403 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8404
8405 - `"claude-sonnet-5-5"`
8406
8407 Efficient model for coding and agents
8408
8409 - `"claude-opus-5-5"`
8410
8411 Powerful intelligence for coding, knowledge work, and long-running agents
8412
8413 - `"claude-fable-5-1"`
8414
8415 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
8416
8417 - `"claude-sonnet-5"`
8418
8419 Efficient model for coding and agents
8420
8421 - `"claude-fable-5"`
8422
8423 Next generation of intelligence for the hardest knowledge work and coding problems
8424
8425 - `"claude-opus-5"`
8426
8427 Powerful intelligence for long-running agents and coding
8428
8429 - `"claude-opus-4-8"`
8430
8431 Powerful intelligence for long-running agents and coding
8432
8433 - `"claude-opus-4-7"`
8434
8435 Powerful intelligence for long-running agents and coding
8436
8437 - `"claude-opus-4-6"`
8438
8439 Powerful intelligence for long-running agents and coding
8440
8441 - `"claude-sonnet-4-6"`
8442
8443 Best combination of speed and intelligence
8444
8445 - `"claude-haiku-4-5"`
8446
8447 Fastest model with near-frontier intelligence
8448
8449 - `"claude-haiku-4-5-20251001"`
8450
8451 Fastest model with near-frontier intelligence
8452
8453 - `"claude-opus-4-5"`
8454
8455 Powerful intelligence for long-running agents and coding
8456
8457 - `"claude-opus-4-5-20251101"`
8458
8459 Powerful intelligence for long-running agents and coding
8460
8461 - `"claude-sonnet-4-5"`
8462
8463 High-performance model for agents and coding
8464
8465 - `"claude-sonnet-4-5-20250929"`
8466
8467 High-performance model for agents and coding
8468
8469 - `string`
8470
8471 - `effort: optional "low" or "medium" or "high" or 2 more or BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or 3 more or null`
8472
8473 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
8474
8475 - `"low"`
8476
8477 Low effort. Favors latency over reasoning depth.
8478
8479 - `"medium"`
8480
8481 Medium effort. Balances latency and reasoning depth.
8482
8483 - `"high"`
8484
8485 High effort. Favors reasoning depth.
8486
8487 - `"xhigh"`
8488
8489 Extra-high effort. Not all models accept this level.
8490
8491 - `"max"`
8492
8493 Maximum effort. Favors reasoning depth over latency.
8494
8495 - `BetaManagedAgentsEffortLow object`
8496
8497 Low effort. Favors latency over reasoning depth.
8498
8499 - `type: "low"`
8500
8501 - `BetaManagedAgentsEffortMedium object`
8502
8503 Medium effort. Balances latency and reasoning depth.
8504
8505 - `type: "medium"`
8506
8507 - `BetaManagedAgentsEffortHigh object`
8508
8509 High effort. Favors reasoning depth.
8510
8511 - `type: "high"`
8512
8513 - `BetaManagedAgentsEffortXhigh object`
8514
8515 Extra-high effort. Not all models accept this level.
8516
8517 - `type: "xhigh"`
8518
8519 - `BetaManagedAgentsEffortMax object`
8520
8521 Maximum effort. Favors reasoning depth over latency.
8522
8523 - `type: "max"`
8524
8525 - `inference_geo: optional string or null`
8526
8527 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo. On update, `model` is whole-object replacement — omitting inference_geo clears it.
8528
8529 - `speed: optional "standard" or "fast" or null`
8530
8531 Inference speed mode. Defaults to `standard`.
8532
8533 - `"standard"`
8534
8535 - `"fast"`
8536
8537### Beta Managed Agents Multiagent Coordinator
8538
8539- `BetaManagedAgentsMultiagentCoordinator object`
8540
8541 Resolved coordinator topology with a concrete agent roster.
8542
8543 - `type: "coordinator"`
8544
8545 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
8546
8547 Agents the coordinator may spawn as session threads, each resolved to a specific version.
8548
8549 - `BetaManagedAgentsAgentReference object`
8550
8551 A resolved agent reference with a concrete version.
8552
8553 - `type: "agent"`
8554
8555 - `id: string`
8556
8557 - `version: number`
8558
8559 format: int32
8560
8561 - `BetaManagedAgentsAdvisor object`
8562
8563 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
8564
8565 - `type: "advisor"`
8566
8567 - `model: string`
8568
8569 The advisor model id.
8570
8571### Beta Managed Agents Multiagent Coordinator Params
8572
8573- `BetaManagedAgentsMultiagentCoordinatorParams object`
8574
8575 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
8576
8577 - `type: "coordinator"`
8578
8579 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
8580
8581 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
8582
8583 - `string`
8584
8585 - `BetaManagedAgentsAgentParams object`
8586
8587 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
8588
8589 - `type: "agent"`
8590
8591 - `id: string`
8592
8593 The `agent` ID.
8594
8595 minLength: 1, maxLength: 128
8596
8597 - `version: optional number`
8598
8599 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
8600
8601 format: int32
8602
8603 - `BetaManagedAgentsMultiagentSelfParams object`
8604
8605 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
8606
8607 - `type: "self"`
8608
8609 - `BetaManagedAgentsAdvisorParams object`
8610
8611 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
8612
8613 - `type: "advisor"`
8614
8615 - `model: string`
8616
8617 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
8618
8619 minLength: 1, maxLength: 256
8620
8621### Beta Managed Agents Multiagent Self Params
8622
8623- `BetaManagedAgentsMultiagentSelfParams object`
8624
8625 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
8626
8627 - `type: "self"`
8628
8629### Beta Managed Agents Read Tool Config
8630
8631- `BetaManagedAgentsReadToolConfig object`
8632
8633 Configuration for the read tool.
8634
8635 - `type: "read"`
8636
8637 - `enabled: boolean`
8638
8639 - `name: "read"`
8640
8641 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8642
8643 Permission policy for tool execution.
8644
8645 - `BetaManagedAgentsAlwaysAllowPolicy object`
8646
8647 Tool calls are automatically approved without user confirmation.
8648
8649 - `type: "always_allow"`
8650
8651 - `BetaManagedAgentsAlwaysAskPolicy object`
8652
8653 Tool calls require user confirmation before execution.
8654
8655 - `type: "always_ask"`
8656
8657 - `BetaManagedAgentsAutoPolicy object`
8658
8659 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8660
8661 - `type: "auto"`
8662
8663### Beta Managed Agents Read Tool Config Params
8664
8665- `BetaManagedAgentsReadToolConfigParams object`
8666
8667 Configuration override for the read tool.
8668
8669 - `type: optional "read"`
8670
8671 - `name: "read"`
8672
8673 Must be "read".
8674
8675 - `enabled: optional boolean or null`
8676
8677 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
8678
8679 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
8680
8681 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
8682
8683 - `BetaManagedAgentsAlwaysAllowPolicy object`
8684
8685 Tool calls are automatically approved without user confirmation.
8686
8687 - `type: "always_allow"`
8688
8689 - `BetaManagedAgentsAlwaysAskPolicy object`
8690
8691 Tool calls require user confirmation before execution.
8692
8693 - `type: "always_ask"`
8694
8695 - `BetaManagedAgentsAutoPolicy object`
8696
8697 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8698
8699 - `type: "auto"`
8700
8701### Beta Managed Agents Session Thread Agent
8702
8703- `BetaManagedAgentsSessionThreadAgent object`
8704
8705 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
8706
8707 - `type: "agent"`
8708
8709 - `id: string`
8710
8711 - `description: string or null`
8712
8713 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
8714
8715 - `type: "url"`
8716
8717 - `name: string`
8718
8719 - `url: string`
8720
8721 - `model: BetaManagedAgentsModelConfig`
8722
8723 Model identifier and configuration.
8724
8725 - `id: BetaManagedAgentsModel`
8726
8727 The model that will power your agent.
8728
8729 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8730
8731 - `"claude-sonnet-5-5"`
8732
8733 Efficient model for coding and agents
8734
8735 - `"claude-opus-5-5"`
8736
8737 Powerful intelligence for coding, knowledge work, and long-running agents
8738
8739 - `"claude-fable-5-1"`
8740
8741 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
8742
8743 - `"claude-sonnet-5"`
8744
8745 Efficient model for coding and agents
8746
8747 - `"claude-fable-5"`
8748
8749 Next generation of intelligence for the hardest knowledge work and coding problems
8750
8751 - `"claude-opus-5"`
8752
8753 Powerful intelligence for long-running agents and coding
8754
8755 - `"claude-opus-4-8"`
8756
8757 Powerful intelligence for long-running agents and coding
8758
8759 - `"claude-opus-4-7"`
8760
8761 Powerful intelligence for long-running agents and coding
8762
8763 - `"claude-opus-4-6"`
8764
8765 Powerful intelligence for long-running agents and coding
8766
8767 - `"claude-sonnet-4-6"`
8768
8769 Best combination of speed and intelligence
8770
8771 - `"claude-haiku-4-5"`
8772
8773 Fastest model with near-frontier intelligence
8774
8775 - `"claude-haiku-4-5-20251001"`
8776
8777 Fastest model with near-frontier intelligence
8778
8779 - `"claude-opus-4-5"`
8780
8781 Powerful intelligence for long-running agents and coding
8782
8783 - `"claude-opus-4-5-20251101"`
8784
8785 Powerful intelligence for long-running agents and coding
8786
8787 - `"claude-sonnet-4-5"`
8788
8789 High-performance model for agents and coding
8790
8791 - `"claude-sonnet-4-5-20250929"`
8792
8793 High-performance model for agents and coding
8794
8795 - `string`
8796
8797 - `effort: optional BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or BetaManagedAgentsEffortHigh or 2 more`
8798
8799 How hard Claude works on each inference call. One of `low`, `medium`, `high`, `xhigh`, `max`. Always present; resolved to the per-model default at save time when not supplied.
8800
8801 - `BetaManagedAgentsEffortLow object`
8802
8803 Low effort. Favors latency over reasoning depth.
8804
8805 - `type: "low"`
8806
8807 - `BetaManagedAgentsEffortMedium object`
8808
8809 Medium effort. Balances latency and reasoning depth.
8810
8811 - `type: "medium"`
8812
8813 - `BetaManagedAgentsEffortHigh object`
8814
8815 High effort. Favors reasoning depth.
8816
8817 - `type: "high"`
8818
8819 - `BetaManagedAgentsEffortXhigh object`
8820
8821 Extra-high effort. Not all models accept this level.
8822
8823 - `type: "xhigh"`
8824
8825 - `BetaManagedAgentsEffortMax object`
8826
8827 Maximum effort. Favors reasoning depth over latency.
8828
8829 - `type: "max"`
8830
8831 - `inference_geo: optional string`
8832
8833 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo.
8834
8835 - `speed: optional "standard" or "fast"`
8836
8837 Inference speed mode. `fast` provides significantly faster output token generation at premium pricing. Defaults to `standard`. Not all models support `fast`; invalid combinations are rejected at create time.
8838
8839 - `"standard"`
8840
8841 - `"fast"`
8842
8843 - `name: string`
8844
8845 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
8846
8847 - `BetaManagedAgentsAnthropicSkill object`
8848
8849 A resolved Anthropic-managed skill.
8850
8851 - `type: "anthropic"`
8852
8853 - `skill_id: string`
8854
8855 - `version: string`
8856
8857 - `BetaManagedAgentsCustomSkill object`
8858
8859 A resolved user-created custom skill.
8860
8861 - `type: "custom"`
8862
8863 - `skill_id: string`
8864
8865 - `version: string`
8866
8867 - `system: string or null`
8868
8869 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
8870
8871 - `BetaManagedAgentsAgentToolset20260401 object`
8872
8873 - `type: "agent_toolset_20260401"`
8874
8875 - `configs: array of BetaManagedAgentsAgentToolConfig`
8876
8877 - `BetaManagedAgentsBashToolConfig object`
8878
8879 Configuration for the bash tool.
8880
8881 - `type: "bash"`
8882
8883 - `enabled: boolean`
8884
8885 - `name: "bash"`
8886
8887 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8888
8889 Permission policy for tool execution.
8890
8891 - `BetaManagedAgentsAlwaysAllowPolicy object`
8892
8893 Tool calls are automatically approved without user confirmation.
8894
8895 - `type: "always_allow"`
8896
8897 - `BetaManagedAgentsAlwaysAskPolicy object`
8898
8899 Tool calls require user confirmation before execution.
8900
8901 - `type: "always_ask"`
8902
8903 - `BetaManagedAgentsAutoPolicy object`
8904
8905 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8906
8907 - `type: "auto"`
8908
8909 - `BetaManagedAgentsEditToolConfig object`
8910
8911 Configuration for the edit tool.
8912
8913 - `type: "edit"`
8914
8915 - `enabled: boolean`
8916
8917 - `name: "edit"`
8918
8919 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8920
8921 Permission policy for tool execution.
8922
8923 - `BetaManagedAgentsAlwaysAllowPolicy object`
8924
8925 Tool calls are automatically approved without user confirmation.
8926
8927 - `BetaManagedAgentsAlwaysAskPolicy object`
8928
8929 Tool calls require user confirmation before execution.
8930
8931 - `BetaManagedAgentsAutoPolicy object`
8932
8933 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8934
8935 - `BetaManagedAgentsReadToolConfig object`
8936
8937 Configuration for the read tool.
8938
8939 - `type: "read"`
8940
8941 - `enabled: boolean`
8942
8943 - `name: "read"`
8944
8945 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8946
8947 Permission policy for tool execution.
8948
8949 - `BetaManagedAgentsAlwaysAllowPolicy object`
8950
8951 Tool calls are automatically approved without user confirmation.
8952
8953 - `BetaManagedAgentsAlwaysAskPolicy object`
8954
8955 Tool calls require user confirmation before execution.
8956
8957 - `BetaManagedAgentsAutoPolicy object`
8958
8959 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8960
8961 - `BetaManagedAgentsWriteToolConfig object`
8962
8963 Configuration for the write tool.
8964
8965 - `type: "write"`
8966
8967 - `enabled: boolean`
8968
8969 - `name: "write"`
8970
8971 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8972
8973 Permission policy for tool execution.
8974
8975 - `BetaManagedAgentsAlwaysAllowPolicy object`
8976
8977 Tool calls are automatically approved without user confirmation.
8978
8979 - `BetaManagedAgentsAlwaysAskPolicy object`
8980
8981 Tool calls require user confirmation before execution.
8982
8983 - `BetaManagedAgentsAutoPolicy object`
8984
8985 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8986
8987 - `BetaManagedAgentsGlobToolConfig object`
8988
8989 Configuration for the glob tool.
8990
8991 - `type: "glob"`
8992
8993 - `enabled: boolean`
8994
8995 - `name: "glob"`
8996
8997 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8998
8999 Permission policy for tool execution.
9000
9001 - `BetaManagedAgentsAlwaysAllowPolicy object`
9002
9003 Tool calls are automatically approved without user confirmation.
9004
9005 - `BetaManagedAgentsAlwaysAskPolicy object`
9006
9007 Tool calls require user confirmation before execution.
9008
9009 - `BetaManagedAgentsAutoPolicy object`
9010
9011 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9012
9013 - `BetaManagedAgentsGrepToolConfig object`
9014
9015 Configuration for the grep tool.
9016
9017 - `type: "grep"`
9018
9019 - `enabled: boolean`
9020
9021 - `name: "grep"`
9022
9023 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9024
9025 Permission policy for tool execution.
9026
9027 - `BetaManagedAgentsAlwaysAllowPolicy object`
9028
9029 Tool calls are automatically approved without user confirmation.
9030
9031 - `BetaManagedAgentsAlwaysAskPolicy object`
9032
9033 Tool calls require user confirmation before execution.
9034
9035 - `BetaManagedAgentsAutoPolicy object`
9036
9037 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9038
9039 - `BetaManagedAgentsWebFetchToolConfig object`
9040
9041 Configuration for the web_fetch tool.
9042
9043 - `type: "web_fetch"`
9044
9045 - `enabled: boolean`
9046
9047 - `name: "web_fetch"`
9048
9049 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9050
9051 Permission policy for tool execution.
9052
9053 - `BetaManagedAgentsAlwaysAllowPolicy object`
9054
9055 Tool calls are automatically approved without user confirmation.
9056
9057 - `BetaManagedAgentsAlwaysAskPolicy object`
9058
9059 Tool calls require user confirmation before execution.
9060
9061 - `BetaManagedAgentsAutoPolicy object`
9062
9063 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9064
9065 - `allowed_domains: optional array of string`
9066
9067 - `blocked_domains: optional array of string`
9068
9069 - `max_content_tokens: optional number or null`
9070
9071 format: int32
9072
9073 - `BetaManagedAgentsWebSearchToolConfig object`
9074
9075 Configuration for the web_search tool.
9076
9077 - `type: "web_search"`
9078
9079 - `enabled: boolean`
9080
9081 - `name: "web_search"`
9082
9083 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9084
9085 Permission policy for tool execution.
9086
9087 - `BetaManagedAgentsAlwaysAllowPolicy object`
9088
9089 Tool calls are automatically approved without user confirmation.
9090
9091 - `BetaManagedAgentsAlwaysAskPolicy object`
9092
9093 Tool calls require user confirmation before execution.
9094
9095 - `BetaManagedAgentsAutoPolicy object`
9096
9097 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9098
9099 - `allowed_domains: optional array of string`
9100
9101 - `blocked_domains: optional array of string`
9102
9103 - `user_location: optional BetaManagedAgentsUserLocation or null`
9104
9105 Approximate user location for search result localization.
9106
9107 - `type: "approximate"`
9108
9109 Location precision. Only "approximate" is supported.
9110
9111 - `city: optional string or null`
9112
9113 City name.
9114
9115 minLength: 1, maxLength: 255
9116
9117 - `country: optional string or null`
9118
9119 Two-letter ISO 3166-1 country code, uppercase.
9120
9121 - `region: optional string or null`
9122
9123 Region or state name.
9124
9125 minLength: 1, maxLength: 255
9126
9127 - `timezone: optional string or null`
9128
9129 IANA timezone identifier, e.g. "America/Los_Angeles".
9130
9131 minLength: 1, maxLength: 255
9132
9133 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
9134
9135 Resolved default configuration for agent tools.
9136
9137 - `enabled: boolean`
9138
9139 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9140
9141 Permission policy for tool execution.
9142
9143 - `BetaManagedAgentsAlwaysAllowPolicy object`
9144
9145 Tool calls are automatically approved without user confirmation.
9146
9147 - `BetaManagedAgentsAlwaysAskPolicy object`
9148
9149 Tool calls require user confirmation before execution.
9150
9151 - `BetaManagedAgentsAutoPolicy object`
9152
9153 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9154
9155 - `BetaManagedAgentsMCPToolset object`
9156
9157 - `type: "mcp_toolset"`
9158
9159 - `configs: array of BetaManagedAgentsMCPToolConfig`
9160
9161 - `enabled: boolean`
9162
9163 - `name: string`
9164
9165 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9166
9167 Permission policy for tool execution.
9168
9169 - `BetaManagedAgentsAlwaysAllowPolicy object`
9170
9171 Tool calls are automatically approved without user confirmation.
9172
9173 - `BetaManagedAgentsAlwaysAskPolicy object`
9174
9175 Tool calls require user confirmation before execution.
9176
9177 - `BetaManagedAgentsAutoPolicy object`
9178
9179 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9180
9181 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
9182
9183 Resolved default configuration for all tools from an MCP server.
9184
9185 - `enabled: boolean`
9186
9187 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9188
9189 Permission policy for tool execution.
9190
9191 - `BetaManagedAgentsAlwaysAllowPolicy object`
9192
9193 Tool calls are automatically approved without user confirmation.
9194
9195 - `BetaManagedAgentsAlwaysAskPolicy object`
9196
9197 Tool calls require user confirmation before execution.
9198
9199 - `BetaManagedAgentsAutoPolicy object`
9200
9201 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9202
9203 - `mcp_server_name: string`
9204
9205 - `BetaManagedAgentsCustomTool object`
9206
9207 A custom tool as returned in API responses.
9208
9209 - `type: "custom"`
9210
9211 - `description: string`
9212
9213 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
9214
9215 JSON Schema for custom tool input parameters.
9216
9217 - `type: "object"`
9218
9219 - `properties: optional map[unknown] or null`
9220
9221 - `required: optional array of string or null`
9222
9223 - `name: string`
9224
9225 - `version: number`
9226
9227 format: int32
9228
9229### Beta Managed Agents Skill Params
9230
9231- `BetaManagedAgentsSkillParams = BetaManagedAgentsAnthropicSkillParams or BetaManagedAgentsCustomSkillParams`
9232
9233 Skill to load in the session container.
9234
9235 - `BetaManagedAgentsAnthropicSkillParams object`
9236
9237 An Anthropic-managed skill.
9238
9239 - `type: "anthropic"`
9240
9241 - `skill_id: string`
9242
9243 Identifier of the Anthropic skill (e.g., "xlsx").
9244
9245 minLength: 1, maxLength: 64
9246
9247 - `version: optional string or null`
9248
9249 Version to pin. Defaults to latest if omitted.
9250
9251 minLength: 1, maxLength: 64
9252
9253 - `BetaManagedAgentsCustomSkillParams object`
9254
9255 A user-created custom skill.
9256
9257 - `type: "custom"`
9258
9259 - `skill_id: string`
9260
9261 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
9262
9263 minLength: 1, maxLength: 64
9264
9265 - `version: optional string or null`
9266
9267 Version to pin. Defaults to latest if omitted.
9268
9269 minLength: 1, maxLength: 64
9270
9271### Beta Managed Agents URL MCP Server Params
9272
9273- `BetaManagedAgentsURLMCPServerParams object`
9274
9275 URL-based MCP server connection.
9276
9277 - `type: "url"`
9278
9279 - `name: string`
9280
9281 Unique name for this server, referenced by mcp_toolset configurations. 1-255 characters.
9282
9283 minLength: 1, maxLength: 255
9284
9285 - `url: string`
9286
9287 Endpoint URL for the MCP server.
9288
9289 maxLength: 2048
9290
9291### Beta Managed Agents User Location
9292
9293- `BetaManagedAgentsUserLocation object`
9294
9295 Approximate user location for search result localization.
9296
9297 - `type: "approximate"`
9298
9299 Location precision. Only "approximate" is supported.
9300
9301 - `city: optional string or null`
9302
9303 City name.
9304
9305 minLength: 1, maxLength: 255
9306
9307 - `country: optional string or null`
9308
9309 Two-letter ISO 3166-1 country code, uppercase.
9310
9311 - `region: optional string or null`
9312
9313 Region or state name.
9314
9315 minLength: 1, maxLength: 255
9316
9317 - `timezone: optional string or null`
9318
9319 IANA timezone identifier, e.g. "America/Los_Angeles".
9320
9321 minLength: 1, maxLength: 255
9322
9323### Beta Managed Agents Web Fetch Tool Config
9324
9325- `BetaManagedAgentsWebFetchToolConfig object`
9326
9327 Configuration for the web_fetch tool.
9328
9329 - `type: "web_fetch"`
9330
9331 - `enabled: boolean`
9332
9333 - `name: "web_fetch"`
9334
9335 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9336
9337 Permission policy for tool execution.
9338
9339 - `BetaManagedAgentsAlwaysAllowPolicy object`
9340
9341 Tool calls are automatically approved without user confirmation.
9342
9343 - `type: "always_allow"`
9344
9345 - `BetaManagedAgentsAlwaysAskPolicy object`
9346
9347 Tool calls require user confirmation before execution.
9348
9349 - `type: "always_ask"`
9350
9351 - `BetaManagedAgentsAutoPolicy object`
9352
9353 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9354
9355 - `type: "auto"`
9356
9357 - `allowed_domains: optional array of string`
9358
9359 - `blocked_domains: optional array of string`
9360
9361 - `max_content_tokens: optional number or null`
9362
9363 format: int32
9364
9365### Beta Managed Agents Web Fetch Tool Config Params
9366
9367- `BetaManagedAgentsWebFetchToolConfigParams object`
9368
9369 Configuration override for the web_fetch tool.
9370
9371 - `type: optional "web_fetch"`
9372
9373 - `name: "web_fetch"`
9374
9375 Must be "web_fetch".
9376
9377 - `allowed_domains: optional array of string`
9378
9379 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
9380
9381 - `blocked_domains: optional array of string`
9382
9383 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
9384
9385 - `enabled: optional boolean or null`
9386
9387 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
9388
9389 - `max_content_tokens: optional number or null`
9390
9391 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
9392
9393 format: int32
9394
9395 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
9396
9397 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
9398
9399 - `BetaManagedAgentsAlwaysAllowPolicy object`
9400
9401 Tool calls are automatically approved without user confirmation.
9402
9403 - `type: "always_allow"`
9404
9405 - `BetaManagedAgentsAlwaysAskPolicy object`
9406
9407 Tool calls require user confirmation before execution.
9408
9409 - `type: "always_ask"`
9410
9411 - `BetaManagedAgentsAutoPolicy object`
9412
9413 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9414
9415 - `type: "auto"`
9416
9417### Beta Managed Agents Web Search Tool Config
9418
9419- `BetaManagedAgentsWebSearchToolConfig object`
9420
9421 Configuration for the web_search tool.
9422
9423 - `type: "web_search"`
9424
9425 - `enabled: boolean`
9426
9427 - `name: "web_search"`
9428
9429 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9430
9431 Permission policy for tool execution.
9432
9433 - `BetaManagedAgentsAlwaysAllowPolicy object`
9434
9435 Tool calls are automatically approved without user confirmation.
9436
9437 - `type: "always_allow"`
9438
9439 - `BetaManagedAgentsAlwaysAskPolicy object`
9440
9441 Tool calls require user confirmation before execution.
9442
9443 - `type: "always_ask"`
9444
9445 - `BetaManagedAgentsAutoPolicy object`
9446
9447 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9448
9449 - `type: "auto"`
9450
9451 - `allowed_domains: optional array of string`
9452
9453 - `blocked_domains: optional array of string`
9454
9455 - `user_location: optional BetaManagedAgentsUserLocation or null`
9456
9457 Approximate user location for search result localization.
9458
9459 - `type: "approximate"`
9460
9461 Location precision. Only "approximate" is supported.
9462
9463 - `city: optional string or null`
9464
9465 City name.
9466
9467 minLength: 1, maxLength: 255
9468
9469 - `country: optional string or null`
9470
9471 Two-letter ISO 3166-1 country code, uppercase.
9472
9473 - `region: optional string or null`
9474
9475 Region or state name.
9476
9477 minLength: 1, maxLength: 255
9478
9479 - `timezone: optional string or null`
9480
9481 IANA timezone identifier, e.g. "America/Los_Angeles".
9482
9483 minLength: 1, maxLength: 255
9484
9485### Beta Managed Agents Web Search Tool Config Params
9486
9487- `BetaManagedAgentsWebSearchToolConfigParams object`
9488
9489 Configuration override for the web_search tool.
9490
9491 - `type: optional "web_search"`
9492
9493 - `name: "web_search"`
9494
9495 Must be "web_search".
9496
9497 - `allowed_domains: optional array of string`
9498
9499 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
9500
9501 - `blocked_domains: optional array of string`
9502
9503 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
9504
9505 - `enabled: optional boolean or null`
9506
9507 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
9508
9509 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
9510
9511 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
9512
9513 - `BetaManagedAgentsAlwaysAllowPolicy object`
9514
9515 Tool calls are automatically approved without user confirmation.
9516
9517 - `type: "always_allow"`
9518
9519 - `BetaManagedAgentsAlwaysAskPolicy object`
9520
9521 Tool calls require user confirmation before execution.
9522
9523 - `type: "always_ask"`
9524
9525 - `BetaManagedAgentsAutoPolicy object`
9526
9527 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9528
9529 - `type: "auto"`
9530
9531 - `user_location: optional BetaManagedAgentsUserLocation or null`
9532
9533 Approximate user location for search result localization.
9534
9535 - `type: "approximate"`
9536
9537 Location precision. Only "approximate" is supported.
9538
9539 - `city: optional string or null`
9540
9541 City name.
9542
9543 minLength: 1, maxLength: 255
9544
9545 - `country: optional string or null`
9546
9547 Two-letter ISO 3166-1 country code, uppercase.
9548
9549 - `region: optional string or null`
9550
9551 Region or state name.
9552
9553 minLength: 1, maxLength: 255
9554
9555 - `timezone: optional string or null`
9556
9557 IANA timezone identifier, e.g. "America/Los_Angeles".
9558
9559 minLength: 1, maxLength: 255
9560
9561### Beta Managed Agents Write Tool Config
9562
9563- `BetaManagedAgentsWriteToolConfig object`
9564
9565 Configuration for the write tool.
9566
9567 - `type: "write"`
9568
9569 - `enabled: boolean`
9570
9571 - `name: "write"`
9572
9573 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9574
9575 Permission policy for tool execution.
9576
9577 - `BetaManagedAgentsAlwaysAllowPolicy object`
9578
9579 Tool calls are automatically approved without user confirmation.
9580
9581 - `type: "always_allow"`
9582
9583 - `BetaManagedAgentsAlwaysAskPolicy object`
9584
9585 Tool calls require user confirmation before execution.
9586
9587 - `type: "always_ask"`
9588
9589 - `BetaManagedAgentsAutoPolicy object`
9590
9591 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9592
9593 - `type: "auto"`
9594
9595### Beta Managed Agents Write Tool Config Params
9596
9597- `BetaManagedAgentsWriteToolConfigParams object`
9598
9599 Configuration override for the write tool.
9600
9601 - `type: optional "write"`
9602
9603 - `name: "write"`
9604
9605 Must be "write".
9606
9607 - `enabled: optional boolean or null`
9608
9609 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
9610
9611 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
9612
9613 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
9614
9615 - `BetaManagedAgentsAlwaysAllowPolicy object`
9616
9617 Tool calls are automatically approved without user confirmation.
9618
9619 - `type: "always_allow"`
9620
9621 - `BetaManagedAgentsAlwaysAskPolicy object`
9622
9623 Tool calls require user confirmation before execution.
9624
9625 - `type: "always_ask"`
9626
9627 - `BetaManagedAgentsAutoPolicy object`
9628
9629 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9630
9631 - `type: "auto"`
9632
9633## Agents › Versions
9634
9635### List Agent Versions
9636
9637**GET** `/v1/agents/{agent_id}/versions`
9638
9639List Agent Versions
9640
9641#### Path parameters
9642
9643- `agent_id: string`
9644
9645 Agent ID to list versions for.
9646
9647#### Query parameters
9648
9649- `limit: optional number`
9650
9651 Maximum results per page. Default 20, maximum 100.
9652
9653 format: int32
9654
9655- `page: optional string`
9656
9657 Opaque pagination cursor.
9658
9659#### Headers
9660
9661- `"anthropic-beta": optional array of AnthropicBeta`
9662
9663 Optional header to specify the beta version(s) you want to use.
9664
83399665 - `string`
83409666
8341### Beta Managed Agents Model Config
8342
8343- `BetaManagedAgentsModelConfig object`
8344
8345 Model identifier and configuration.
8346
8347 - `id: BetaManagedAgentsModel`
8348
8349 The model that will power your agent.
8350
8351 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8352
8353 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
9667 - `"message-batches-2024-09-24"`
9668
9669 - `"prompt-caching-2024-07-31"`
9670
9671 - `"computer-use-2024-10-22"`
9672
9673 - `"computer-use-2025-01-24"`
9674
9675 - `"pdfs-2024-09-25"`
9676
9677 - `"token-counting-2024-11-01"`
9678
9679 - `"token-efficient-tools-2025-02-19"`
9680
9681 - `"output-128k-2025-02-19"`
9682
9683 - `"files-api-2025-04-14"`
9684
9685 - `"mcp-client-2025-04-04"`
9686
9687 - `"mcp-client-2025-11-20"`
9688
9689 - `"dev-full-thinking-2025-05-14"`
9690
9691 - `"interleaved-thinking-2025-05-14"`
9692
9693 - `"code-execution-2025-05-22"`
9694
9695 - `"extended-cache-ttl-2025-04-11"`
9696
9697 - `"context-1m-2025-08-07"`
9698
9699 - `"context-management-2025-06-27"`
9700
9701 - `"model-context-window-exceeded-2025-08-26"`
9702
9703 - `"skills-2025-10-02"`
9704
9705 - `"fast-mode-2026-02-01"`
9706
9707 - `"output-300k-2026-03-24"`
9708
9709 - `"user-profiles-2026-03-24"`
9710
9711 - `"user-profiles-2026-08-18"`
9712
9713 - `"user-profiles-2026-09-04"`
9714
9715 - `"advisor-tool-2026-03-01"`
9716
9717 - `"managed-agents-2026-04-01"`
9718
9719 - `"cache-diagnosis-2026-04-07"`
9720
9721 - `"dreaming-2026-04-21"`
9722
9723 - `"thinking-token-count-2026-05-13"`
9724
9725 - `"server-side-fallback-2026-06-01"`
9726
9727 - `"server-side-fallback-2026-07-01"`
9728
9729 - `"fallback-credit-2026-06-01"`
9730
9731 - `"fallback-credit-2026-07-01"`
9732
9733 - `"agent-memory-2026-07-22"`
9734
9735 - `"mid-conversation-tool-changes-2026-07-01"`
9736
9737 - `"compact-2026-01-12"`
9738
9739 - `"computer-use-2025-11-24"`
9740
9741 - `"mcp-tunnels-2026-06-22"`
9742
9743 - `"structured-outputs-2025-11-13"`
9744
9745 - `"task-budgets-2026-03-13"`
9746
9747 - `"thinking-display-updates-2026-08-18"`
9748
9749 - `"ce-user-management-2026-07-13"`
9750
9751 - `"mid-conversation-output-config-2026-07-01"`
9752
9753 - `"thinking-binding-controls-2026-08-01"`
9754
9755 - `"mid-conversation-system-clear-at-2026-08-21"`
9756
9757 - `"compact-2026-09-04"`
9758
9759 - `"inline-tools-2026-09-15"`
9760
9761 - `"mcp-client-2026-09-15"`
9762
9763- `"anthropic-workspace-id": optional string`
9764
9765 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
9766
9767 Only needed for credentials that can act on more than one Workspace. A credential that belongs to a specific Workspace may omit it; if sent, it must match that Workspace.
9768
9769#### Returns
9770
9771- `data: array of BetaManagedAgentsAgent`
9772
9773 Agent versions.
9774
9775 - `type: "agent"`
9776
9777 - `id: string`
9778
9779 - `archived_at: string or null`
9780
9781 When the agent was archived. Null if not archived.
9782
9783 format: date-time
9784
9785 - `created_at: string`
9786
9787 A timestamp in RFC 3339 format
9788
9789 format: date-time
9790
9791 - `description: string or null`
9792
9793 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
9794
9795 - `type: "url"`
9796
9797 - `name: string`
9798
9799 - `url: string`
9800
9801 - `metadata: map[string]`
9802
9803 - `model: BetaManagedAgentsModelConfig`
9804
9805 Model identifier and configuration.
9806
9807 - `id: BetaManagedAgentsModel`
83549808
83559809 The model that will power your agent.
83569810
from line 9873
84199873 - `"claude-sonnet-4-5-20250929"`
84209874
84219875 High-performance model for agents and coding
8422
8423 - `string`
8424
8425 - `effort: optional BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or BetaManagedAgentsEffortHigh or 2 more`
8426
8427 How hard Claude works on each inference call. One of `low`, `medium`, `high`, `xhigh`, `max`. Always present; resolved to the per-model default at save time when not supplied.
8428
8429 - `BetaManagedAgentsEffortLow object`
8430
8431 Low effort. Favors latency over reasoning depth.
8432
8433 - `type: "low"`
8434
8435 - `BetaManagedAgentsEffortMedium object`
8436
8437 Medium effort. Balances latency and reasoning depth.
8438
8439 - `type: "medium"`
8440
8441 - `BetaManagedAgentsEffortHigh object`
8442
8443 High effort. Favors reasoning depth.
8444
8445 - `type: "high"`
8446
8447 - `BetaManagedAgentsEffortXhigh object`
8448
8449 Extra-high effort. Not all models accept this level.
8450
8451 - `type: "xhigh"`
8452
8453 - `BetaManagedAgentsEffortMax object`
8454
8455 Maximum effort. Favors reasoning depth over latency.
8456
8457 - `type: "max"`
8458
8459 - `inference_geo: optional string`
8460
8461 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo.
8462
8463 - `speed: optional "standard" or "fast"`
8464
8465 Inference speed mode. `fast` provides significantly faster output token generation at premium pricing. Defaults to `standard`. Not all models support `fast`; invalid combinations are rejected at create time.
8466
8467 - `"standard"`
8468
8469 - `"fast"`
8470
8471### Beta Managed Agents Model Config Params
8472
8473- `BetaManagedAgentsModelConfigParams object`
8474
8475 An object that defines additional configuration control over model use
8476
8477 - `id: BetaManagedAgentsModel`
8478
8479 The model that will power your agent.
8480
8481 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8482
8483 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
8484
8485 The model that will power your agent.
8486
8487 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8488
8489 - `"claude-sonnet-5-5"`
8490
8491 Efficient model for coding and agents
8492
8493 - `"claude-opus-5-5"`
8494
8495 Powerful intelligence for coding, knowledge work, and long-running agents
8496
8497 - `"claude-fable-5-1"`
8498
8499 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
8500
8501 - `"claude-sonnet-5"`
8502
8503 Efficient model for coding and agents
8504
8505 - `"claude-fable-5"`
8506
8507 Next generation of intelligence for the hardest knowledge work and coding problems
8508
8509 - `"claude-opus-5"`
8510
8511 Powerful intelligence for long-running agents and coding
8512
8513 - `"claude-opus-4-8"`
8514
8515 Powerful intelligence for long-running agents and coding
8516
8517 - `"claude-opus-4-7"`
8518
8519 Powerful intelligence for long-running agents and coding
8520
8521 - `"claude-opus-4-6"`
8522
8523 Powerful intelligence for long-running agents and coding
8524
8525 - `"claude-sonnet-4-6"`
8526
8527 Best combination of speed and intelligence
8528
8529 - `"claude-haiku-4-5"`
8530
8531 Fastest model with near-frontier intelligence
8532
8533 - `"claude-haiku-4-5-20251001"`
8534
8535 Fastest model with near-frontier intelligence
8536
8537 - `"claude-opus-4-5"`
8538
8539 Powerful intelligence for long-running agents and coding
8540
8541 - `"claude-opus-4-5-20251101"`
8542
8543 Powerful intelligence for long-running agents and coding
8544
8545 - `"claude-sonnet-4-5"`
8546
8547 High-performance model for agents and coding
8548
8549 - `"claude-sonnet-4-5-20250929"`
8550
8551 High-performance model for agents and coding
8552
8553 - `string`
8554
8555 - `effort: optional "low" or "medium" or "high" or 2 more or BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or 3 more or null`
8556
8557 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
8558
8559 - `BetaManagedAgentsEffortLevel = "low" or "medium" or "high" or 2 more`
8560
8561 How hard Claude works on each turn. Higher levels favor reasoning depth over latency. Not all models accept every level; invalid combinations are rejected at create time.
8562
8563 - `"low"`
8564
8565 Low effort. Favors latency over reasoning depth.
8566
8567 - `"medium"`
8568
8569 Medium effort. Balances latency and reasoning depth.
8570
8571 - `"high"`
8572
8573 High effort. Favors reasoning depth.
8574
8575 - `"xhigh"`
8576
8577 Extra-high effort. Not all models accept this level.
8578
8579 - `"max"`
8580
8581 Maximum effort. Favors reasoning depth over latency.
8582
8583 - `BetaManagedAgentsEffortLow object`
8584
8585 Low effort. Favors latency over reasoning depth.
8586
8587 - `type: "low"`
8588
8589 - `BetaManagedAgentsEffortMedium object`
8590
8591 Medium effort. Balances latency and reasoning depth.
8592
8593 - `type: "medium"`
8594
8595 - `BetaManagedAgentsEffortHigh object`
8596
8597 High effort. Favors reasoning depth.
8598
8599 - `type: "high"`
8600
8601 - `BetaManagedAgentsEffortXhigh object`
8602
8603 Extra-high effort. Not all models accept this level.
8604
8605 - `type: "xhigh"`
8606
8607 - `BetaManagedAgentsEffortMax object`
8608
8609 Maximum effort. Favors reasoning depth over latency.
8610
8611 - `type: "max"`
8612
8613 - `inference_geo: optional string or null`
8614
8615 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo. On update, `model` is whole-object replacement — omitting inference_geo clears it.
8616
8617 - `speed: optional "standard" or "fast" or null`
8618
8619 Inference speed mode. Defaults to `standard`.
8620
8621 - `"standard"`
8622
8623 - `"fast"`
8624
8625### Beta Managed Agents Multiagent Coordinator
8626
8627- `BetaManagedAgentsMultiagentCoordinator object`
8628
8629 Resolved coordinator topology with a concrete agent roster.
8630
8631 - `type: "coordinator"`
8632
8633 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
8634
8635 Agents the coordinator may spawn as session threads, each resolved to a specific version.
8636
8637 - `BetaManagedAgentsAgentReference object`
8638
8639 A resolved agent reference with a concrete version.
8640
8641 - `type: "agent"`
8642
8643 - `id: string`
8644
8645 - `version: number`
8646
8647 format: int32
8648
8649 - `BetaManagedAgentsAdvisor object`
8650
8651 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
8652
8653 - `type: "advisor"`
8654
8655 - `model: string`
8656
8657 The advisor model id.
8658
8659### Beta Managed Agents Multiagent Coordinator Params
8660
8661- `BetaManagedAgentsMultiagentCoordinatorParams object`
8662
8663 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
8664
8665 - `type: "coordinator"`
8666
8667 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
8668
8669 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
8670
8671 - `string`
8672
8673 - `BetaManagedAgentsAgentParams object`
8674
8675 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
8676
8677 - `type: "agent"`
8678
8679 - `id: string`
8680
8681 The `agent` ID.
8682
8683 minLength: 1, maxLength: 128
8684
8685 - `version: optional number`
8686
8687 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
8688
8689 format: int32
8690
8691 - `BetaManagedAgentsMultiagentSelfParams object`
8692
8693 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
8694
8695 - `type: "self"`
8696
8697 - `BetaManagedAgentsAdvisorParams object`
8698
8699 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
8700
8701 - `type: "advisor"`
8702
8703 - `model: string`
8704
8705 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
8706
8707 minLength: 1, maxLength: 256
8708
8709### Beta Managed Agents Multiagent Self Params
8710
8711- `BetaManagedAgentsMultiagentSelfParams object`
8712
8713 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
8714
8715 - `type: "self"`
8716
8717### Beta Managed Agents Read Tool Config
8718
8719- `BetaManagedAgentsReadToolConfig object`
8720
8721 Configuration for the read tool.
8722
8723 - `type: "read"`
8724
8725 - `enabled: boolean`
8726
8727 - `name: "read"`
8728
8729 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8730
8731 Permission policy for tool execution.
8732
8733 - `BetaManagedAgentsAlwaysAllowPolicy object`
8734
8735 Tool calls are automatically approved without user confirmation.
8736
8737 - `type: "always_allow"`
8738
8739 - `BetaManagedAgentsAlwaysAskPolicy object`
8740
8741 Tool calls require user confirmation before execution.
8742
8743 - `type: "always_ask"`
8744
8745 - `BetaManagedAgentsAutoPolicy object`
8746
8747 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8748
8749 - `type: "auto"`
8750
8751### Beta Managed Agents Read Tool Config Params
8752
8753- `BetaManagedAgentsReadToolConfigParams object`
8754
8755 Configuration override for the read tool.
8756
8757 - `type: optional "read"`
8758
8759 - `name: "read"`
8760
8761 Must be "read".
8762
8763 - `enabled: optional boolean or null`
8764
8765 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
8766
8767 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
8768
8769 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
8770
8771 - `BetaManagedAgentsAlwaysAllowPolicy object`
8772
8773 Tool calls are automatically approved without user confirmation.
8774
8775 - `type: "always_allow"`
8776
8777 - `BetaManagedAgentsAlwaysAskPolicy object`
8778
8779 Tool calls require user confirmation before execution.
8780
8781 - `type: "always_ask"`
8782
8783 - `BetaManagedAgentsAutoPolicy object`
8784
8785 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
8786
8787 - `type: "auto"`
8788
8789### Beta Managed Agents Session Thread Agent
8790
8791- `BetaManagedAgentsSessionThreadAgent object`
8792
8793 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
8794
8795 - `type: "agent"`
8796
8797 - `id: string`
8798
8799 - `description: string or null`
8800
8801 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
8802
8803 - `type: "url"`
8804
8805 - `name: string`
8806
8807 - `url: string`
8808
8809 - `model: BetaManagedAgentsModelConfig`
8810
8811 Model identifier and configuration.
8812
8813 - `id: BetaManagedAgentsModel`
8814
8815 The model that will power your agent.
8816
8817 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8818
8819 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
8820
8821 The model that will power your agent.
8822
8823 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
8824
8825 - `"claude-sonnet-5-5"`
8826
8827 Efficient model for coding and agents
8828
8829 - `"claude-opus-5-5"`
8830
8831 Powerful intelligence for coding, knowledge work, and long-running agents
8832
8833 - `"claude-fable-5-1"`
8834
8835 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
8836
8837 - `"claude-sonnet-5"`
8838
8839 Efficient model for coding and agents
8840
8841 - `"claude-fable-5"`
8842
8843 Next generation of intelligence for the hardest knowledge work and coding problems
8844
8845 - `"claude-opus-5"`
8846
8847 Powerful intelligence for long-running agents and coding
8848
8849 - `"claude-opus-4-8"`
8850
8851 Powerful intelligence for long-running agents and coding
8852
8853 - `"claude-opus-4-7"`
8854
8855 Powerful intelligence for long-running agents and coding
8856
8857 - `"claude-opus-4-6"`
8858
8859 Powerful intelligence for long-running agents and coding
8860
8861 - `"claude-sonnet-4-6"`
8862
8863 Best combination of speed and intelligence
8864
8865 - `"claude-haiku-4-5"`
8866
8867 Fastest model with near-frontier intelligence
8868
8869 - `"claude-haiku-4-5-20251001"`
8870
8871 Fastest model with near-frontier intelligence
8872
8873 - `"claude-opus-4-5"`
8874
8875 Powerful intelligence for long-running agents and coding
8876
8877 - `"claude-opus-4-5-20251101"`
8878
8879 Powerful intelligence for long-running agents and coding
8880
8881 - `"claude-sonnet-4-5"`
8882
8883 High-performance model for agents and coding
8884
8885 - `"claude-sonnet-4-5-20250929"`
8886
8887 High-performance model for agents and coding
8888
8889 - `string`
8890
8891 - `effort: optional BetaManagedAgentsEffortLow or BetaManagedAgentsEffortMedium or BetaManagedAgentsEffortHigh or 2 more`
8892
8893 How hard Claude works on each inference call. One of `low`, `medium`, `high`, `xhigh`, `max`. Always present; resolved to the per-model default at save time when not supplied.
8894
8895 - `BetaManagedAgentsEffortLow object`
8896
8897 Low effort. Favors latency over reasoning depth.
8898
8899 - `type: "low"`
8900
8901 - `BetaManagedAgentsEffortMedium object`
8902
8903 Medium effort. Balances latency and reasoning depth.
8904
8905 - `type: "medium"`
8906
8907 - `BetaManagedAgentsEffortHigh object`
8908
8909 High effort. Favors reasoning depth.
8910
8911 - `type: "high"`
8912
8913 - `BetaManagedAgentsEffortXhigh object`
8914
8915 Extra-high effort. Not all models accept this level.
8916
8917 - `type: "xhigh"`
8918
8919 - `BetaManagedAgentsEffortMax object`
8920
8921 Maximum effort. Favors reasoning depth over latency.
8922
8923 - `type: "max"`
8924
8925 - `inference_geo: optional string`
8926
8927 Geographic region for model inference. When unset, requests fall through to the workspace's default_inference_geo.
8928
8929 - `speed: optional "standard" or "fast"`
8930
8931 Inference speed mode. `fast` provides significantly faster output token generation at premium pricing. Defaults to `standard`. Not all models support `fast`; invalid combinations are rejected at create time.
8932
8933 - `"standard"`
8934
8935 - `"fast"`
8936
8937 - `name: string`
8938
8939 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
8940
8941 - `BetaManagedAgentsAnthropicSkill object`
8942
8943 A resolved Anthropic-managed skill.
8944
8945 - `type: "anthropic"`
8946
8947 - `skill_id: string`
8948
8949 - `version: string`
8950
8951 - `BetaManagedAgentsCustomSkill object`
8952
8953 A resolved user-created custom skill.
8954
8955 - `type: "custom"`
8956
8957 - `skill_id: string`
8958
8959 - `version: string`
8960
8961 - `system: string or null`
8962
8963 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
8964
8965 - `BetaManagedAgentsAgentToolset20260401 object`
8966
8967 - `type: "agent_toolset_20260401"`
8968
8969 - `configs: array of BetaManagedAgentsAgentToolConfig`
8970
8971 - `BetaManagedAgentsBashToolConfig object`
8972
8973 Configuration for the bash tool.
8974
8975 - `type: "bash"`
8976
8977 - `enabled: boolean`
8978
8979 - `name: "bash"`
8980
8981 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
8982
8983 Permission policy for tool execution.
8984
8985 - `BetaManagedAgentsAlwaysAllowPolicy object`
8986
8987 Tool calls are automatically approved without user confirmation.
8988
8989 - `type: "always_allow"`
8990
8991 - `BetaManagedAgentsAlwaysAskPolicy object`
8992
8993 Tool calls require user confirmation before execution.
8994
8995 - `type: "always_ask"`
8996
8997 - `BetaManagedAgentsAutoPolicy object`
8998
8999 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9000
9001 - `type: "auto"`
9002
9003 - `BetaManagedAgentsEditToolConfig object`
9004
9005 Configuration for the edit tool.
9006
9007 - `type: "edit"`
9008
9009 - `enabled: boolean`
9010
9011 - `name: "edit"`
9012
9013 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9014
9015 Permission policy for tool execution.
9016
9017 - `BetaManagedAgentsAlwaysAllowPolicy object`
9018
9019 Tool calls are automatically approved without user confirmation.
9020
9021 - `BetaManagedAgentsAlwaysAskPolicy object`
9022
9023 Tool calls require user confirmation before execution.
9024
9025 - `BetaManagedAgentsAutoPolicy object`
9026
9027 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9028
9029 - `BetaManagedAgentsReadToolConfig object`
9030
9031 Configuration for the read tool.
9032
9033 - `type: "read"`
9034
9035 - `enabled: boolean`
9036
9037 - `name: "read"`
9038
9039 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9040
9041 Permission policy for tool execution.
9042
9043 - `BetaManagedAgentsAlwaysAllowPolicy object`
9044
9045 Tool calls are automatically approved without user confirmation.
9046
9047 - `BetaManagedAgentsAlwaysAskPolicy object`
9048
9049 Tool calls require user confirmation before execution.
9050
9051 - `BetaManagedAgentsAutoPolicy object`
9052
9053 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9054
9055 - `BetaManagedAgentsWriteToolConfig object`
9056
9057 Configuration for the write tool.
9058
9059 - `type: "write"`
9060
9061 - `enabled: boolean`
9062
9063 - `name: "write"`
9064
9065 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9066
9067 Permission policy for tool execution.
9068
9069 - `BetaManagedAgentsAlwaysAllowPolicy object`
9070
9071 Tool calls are automatically approved without user confirmation.
9072
9073 - `BetaManagedAgentsAlwaysAskPolicy object`
9074
9075 Tool calls require user confirmation before execution.
9076
9077 - `BetaManagedAgentsAutoPolicy object`
9078
9079 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9080
9081 - `BetaManagedAgentsGlobToolConfig object`
9082
9083 Configuration for the glob tool.
9084
9085 - `type: "glob"`
9086
9087 - `enabled: boolean`
9088
9089 - `name: "glob"`
9090
9091 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9092
9093 Permission policy for tool execution.
9094
9095 - `BetaManagedAgentsAlwaysAllowPolicy object`
9096
9097 Tool calls are automatically approved without user confirmation.
9098
9099 - `BetaManagedAgentsAlwaysAskPolicy object`
9100
9101 Tool calls require user confirmation before execution.
9102
9103 - `BetaManagedAgentsAutoPolicy object`
9104
9105 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9106
9107 - `BetaManagedAgentsGrepToolConfig object`
9108
9109 Configuration for the grep tool.
9110
9111 - `type: "grep"`
9112
9113 - `enabled: boolean`
9114
9115 - `name: "grep"`
9116
9117 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9118
9119 Permission policy for tool execution.
9120
9121 - `BetaManagedAgentsAlwaysAllowPolicy object`
9122
9123 Tool calls are automatically approved without user confirmation.
9124
9125 - `BetaManagedAgentsAlwaysAskPolicy object`
9126
9127 Tool calls require user confirmation before execution.
9128
9129 - `BetaManagedAgentsAutoPolicy object`
9130
9131 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9132
9133 - `BetaManagedAgentsWebFetchToolConfig object`
9134
9135 Configuration for the web_fetch tool.
9136
9137 - `type: "web_fetch"`
9138
9139 - `enabled: boolean`
9140
9141 - `name: "web_fetch"`
9142
9143 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9144
9145 Permission policy for tool execution.
9146
9147 - `BetaManagedAgentsAlwaysAllowPolicy object`
9148
9149 Tool calls are automatically approved without user confirmation.
9150
9151 - `BetaManagedAgentsAlwaysAskPolicy object`
9152
9153 Tool calls require user confirmation before execution.
9154
9155 - `BetaManagedAgentsAutoPolicy object`
9156
9157 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9158
9159 - `allowed_domains: optional array of string`
9160
9161 - `blocked_domains: optional array of string`
9162
9163 - `max_content_tokens: optional number or null`
9164
9165 format: int32
9166
9167 - `BetaManagedAgentsWebSearchToolConfig object`
9168
9169 Configuration for the web_search tool.
9170
9171 - `type: "web_search"`
9172
9173 - `enabled: boolean`
9174
9175 - `name: "web_search"`
9176
9177 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9178
9179 Permission policy for tool execution.
9180
9181 - `BetaManagedAgentsAlwaysAllowPolicy object`
9182
9183 Tool calls are automatically approved without user confirmation.
9184
9185 - `BetaManagedAgentsAlwaysAskPolicy object`
9186
9187 Tool calls require user confirmation before execution.
9188
9189 - `BetaManagedAgentsAutoPolicy object`
9190
9191 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9192
9193 - `allowed_domains: optional array of string`
9194
9195 - `blocked_domains: optional array of string`
9196
9197 - `user_location: optional BetaManagedAgentsUserLocation or null`
9198
9199 Approximate user location for search result localization.
9200
9201 - `type: "approximate"`
9202
9203 Location precision. Only "approximate" is supported.
9204
9205 - `city: optional string or null`
9206
9207 City name.
9208
9209 minLength: 1, maxLength: 255
9210
9211 - `country: optional string or null`
9212
9213 Two-letter ISO 3166-1 country code, uppercase.
9214
9215 - `region: optional string or null`
9216
9217 Region or state name.
9218
9219 minLength: 1, maxLength: 255
9220
9221 - `timezone: optional string or null`
9222
9223 IANA timezone identifier, e.g. "America/Los_Angeles".
9224
9225 minLength: 1, maxLength: 255
9226
9227 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
9228
9229 Resolved default configuration for agent tools.
9230
9231 - `enabled: boolean`
9232
9233 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9234
9235 Permission policy for tool execution.
9236
9237 - `BetaManagedAgentsAlwaysAllowPolicy object`
9238
9239 Tool calls are automatically approved without user confirmation.
9240
9241 - `BetaManagedAgentsAlwaysAskPolicy object`
9242
9243 Tool calls require user confirmation before execution.
9244
9245 - `BetaManagedAgentsAutoPolicy object`
9246
9247 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9248
9249 - `BetaManagedAgentsMCPToolset object`
9250
9251 - `type: "mcp_toolset"`
9252
9253 - `configs: array of BetaManagedAgentsMCPToolConfig`
9254
9255 - `enabled: boolean`
9256
9257 - `name: string`
9258
9259 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9260
9261 Permission policy for tool execution.
9262
9263 - `BetaManagedAgentsAlwaysAllowPolicy object`
9264
9265 Tool calls are automatically approved without user confirmation.
9266
9267 - `BetaManagedAgentsAlwaysAskPolicy object`
9268
9269 Tool calls require user confirmation before execution.
9270
9271 - `BetaManagedAgentsAutoPolicy object`
9272
9273 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9274
9275 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
9276
9277 Resolved default configuration for all tools from an MCP server.
9278
9279 - `enabled: boolean`
9280
9281 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9282
9283 Permission policy for tool execution.
9284
9285 - `BetaManagedAgentsAlwaysAllowPolicy object`
9286
9287 Tool calls are automatically approved without user confirmation.
9288
9289 - `BetaManagedAgentsAlwaysAskPolicy object`
9290
9291 Tool calls require user confirmation before execution.
9292
9293 - `BetaManagedAgentsAutoPolicy object`
9294
9295 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9296
9297 - `mcp_server_name: string`
9298
9299 - `BetaManagedAgentsCustomTool object`
9300
9301 A custom tool as returned in API responses.
9302
9303 - `type: "custom"`
9304
9305 - `description: string`
9306
9307 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
9308
9309 JSON Schema for custom tool input parameters.
9310
9311 - `type: "object"`
9312
9313 - `properties: optional map[unknown] or null`
9314
9315 - `required: optional array of string or null`
9316
9317 - `name: string`
9318
9319 - `version: number`
9320
9321 format: int32
9322
9323### Beta Managed Agents Skill Params
9324
9325- `BetaManagedAgentsSkillParams = BetaManagedAgentsAnthropicSkillParams or BetaManagedAgentsCustomSkillParams`
9326
9327 Skill to load in the session container.
9328
9329 - `BetaManagedAgentsAnthropicSkillParams object`
9330
9331 An Anthropic-managed skill.
9332
9333 - `type: "anthropic"`
9334
9335 - `skill_id: string`
9336
9337 Identifier of the Anthropic skill (e.g., "xlsx").
9338
9339 minLength: 1, maxLength: 64
9340
9341 - `version: optional string or null`
9342
9343 Version to pin. Defaults to latest if omitted.
9344
9345 minLength: 1, maxLength: 64
9346
9347 - `BetaManagedAgentsCustomSkillParams object`
9348
9349 A user-created custom skill.
9350
9351 - `type: "custom"`
9352
9353 - `skill_id: string`
9354
9355 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
9356
9357 minLength: 1, maxLength: 64
9358
9359 - `version: optional string or null`
9360
9361 Version to pin. Defaults to latest if omitted.
9362
9363 minLength: 1, maxLength: 64
9364
9365### Beta Managed Agents URL MCP Server Params
9366
9367- `BetaManagedAgentsURLMCPServerParams object`
9368
9369 URL-based MCP server connection.
9370
9371 - `type: "url"`
9372
9373 - `name: string`
9374
9375 Unique name for this server, referenced by mcp_toolset configurations. 1-255 characters.
9376
9377 minLength: 1, maxLength: 255
9378
9379 - `url: string`
9380
9381 Endpoint URL for the MCP server.
9382
9383 maxLength: 2048
9384
9385### Beta Managed Agents User Location
9386
9387- `BetaManagedAgentsUserLocation object`
9388
9389 Approximate user location for search result localization.
9390
9391 - `type: "approximate"`
9392
9393 Location precision. Only "approximate" is supported.
9394
9395 - `city: optional string or null`
9396
9397 City name.
9398
9399 minLength: 1, maxLength: 255
9400
9401 - `country: optional string or null`
9402
9403 Two-letter ISO 3166-1 country code, uppercase.
9404
9405 - `region: optional string or null`
9406
9407 Region or state name.
9408
9409 minLength: 1, maxLength: 255
9410
9411 - `timezone: optional string or null`
9412
9413 IANA timezone identifier, e.g. "America/Los_Angeles".
9414
9415 minLength: 1, maxLength: 255
9416
9417### Beta Managed Agents Web Fetch Tool Config
9418
9419- `BetaManagedAgentsWebFetchToolConfig object`
9420
9421 Configuration for the web_fetch tool.
9422
9423 - `type: "web_fetch"`
9424
9425 - `enabled: boolean`
9426
9427 - `name: "web_fetch"`
9428
9429 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9430
9431 Permission policy for tool execution.
9432
9433 - `BetaManagedAgentsAlwaysAllowPolicy object`
9434
9435 Tool calls are automatically approved without user confirmation.
9436
9437 - `type: "always_allow"`
9438
9439 - `BetaManagedAgentsAlwaysAskPolicy object`
9440
9441 Tool calls require user confirmation before execution.
9442
9443 - `type: "always_ask"`
9444
9445 - `BetaManagedAgentsAutoPolicy object`
9446
9447 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9448
9449 - `type: "auto"`
9450
9451 - `allowed_domains: optional array of string`
9452
9453 - `blocked_domains: optional array of string`
9454
9455 - `max_content_tokens: optional number or null`
9456
9457 format: int32
9458
9459### Beta Managed Agents Web Fetch Tool Config Params
9460
9461- `BetaManagedAgentsWebFetchToolConfigParams object`
9462
9463 Configuration override for the web_fetch tool.
9464
9465 - `type: optional "web_fetch"`
9466
9467 - `name: "web_fetch"`
9468
9469 Must be "web_fetch".
9470
9471 - `allowed_domains: optional array of string`
9472
9473 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
9474
9475 - `blocked_domains: optional array of string`
9476
9477 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
9478
9479 - `enabled: optional boolean or null`
9480
9481 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
9482
9483 - `max_content_tokens: optional number or null`
9484
9485 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
9486
9487 format: int32
9488
9489 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
9490
9491 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
9492
9493 - `BetaManagedAgentsAlwaysAllowPolicy object`
9494
9495 Tool calls are automatically approved without user confirmation.
9496
9497 - `type: "always_allow"`
9498
9499 - `BetaManagedAgentsAlwaysAskPolicy object`
9500
9501 Tool calls require user confirmation before execution.
9502
9503 - `type: "always_ask"`
9504
9505 - `BetaManagedAgentsAutoPolicy object`
9506
9507 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9508
9509 - `type: "auto"`
9510
9511### Beta Managed Agents Web Search Tool Config
9512
9513- `BetaManagedAgentsWebSearchToolConfig object`
9514
9515 Configuration for the web_search tool.
9516
9517 - `type: "web_search"`
9518
9519 - `enabled: boolean`
9520
9521 - `name: "web_search"`
9522
9523 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9524
9525 Permission policy for tool execution.
9526
9527 - `BetaManagedAgentsAlwaysAllowPolicy object`
9528
9529 Tool calls are automatically approved without user confirmation.
9530
9531 - `type: "always_allow"`
9532
9533 - `BetaManagedAgentsAlwaysAskPolicy object`
9534
9535 Tool calls require user confirmation before execution.
9536
9537 - `type: "always_ask"`
9538
9539 - `BetaManagedAgentsAutoPolicy object`
9540
9541 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9542
9543 - `type: "auto"`
9544
9545 - `allowed_domains: optional array of string`
9546
9547 - `blocked_domains: optional array of string`
9548
9549 - `user_location: optional BetaManagedAgentsUserLocation or null`
9550
9551 Approximate user location for search result localization.
9552
9553 - `type: "approximate"`
9554
9555 Location precision. Only "approximate" is supported.
9556
9557 - `city: optional string or null`
9558
9559 City name.
9560
9561 minLength: 1, maxLength: 255
9562
9563 - `country: optional string or null`
9564
9565 Two-letter ISO 3166-1 country code, uppercase.
9566
9567 - `region: optional string or null`
9568
9569 Region or state name.
9570
9571 minLength: 1, maxLength: 255
9572
9573 - `timezone: optional string or null`
9574
9575 IANA timezone identifier, e.g. "America/Los_Angeles".
9576
9577 minLength: 1, maxLength: 255
9578
9579### Beta Managed Agents Web Search Tool Config Params
9580
9581- `BetaManagedAgentsWebSearchToolConfigParams object`
9582
9583 Configuration override for the web_search tool.
9584
9585 - `type: optional "web_search"`
9586
9587 - `name: "web_search"`
9588
9589 Must be "web_search".
9590
9591 - `allowed_domains: optional array of string`
9592
9593 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
9594
9595 - `blocked_domains: optional array of string`
9596
9597 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
9598
9599 - `enabled: optional boolean or null`
9600
9601 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
9602
9603 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
9604
9605 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
9606
9607 - `BetaManagedAgentsAlwaysAllowPolicy object`
9608
9609 Tool calls are automatically approved without user confirmation.
9610
9611 - `type: "always_allow"`
9612
9613 - `BetaManagedAgentsAlwaysAskPolicy object`
9614
9615 Tool calls require user confirmation before execution.
9616
9617 - `type: "always_ask"`
9618
9619 - `BetaManagedAgentsAutoPolicy object`
9620
9621 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9622
9623 - `type: "auto"`
9624
9625 - `user_location: optional BetaManagedAgentsUserLocation or null`
9626
9627 Approximate user location for search result localization.
9628
9629 - `type: "approximate"`
9630
9631 Location precision. Only "approximate" is supported.
9632
9633 - `city: optional string or null`
9634
9635 City name.
9636
9637 minLength: 1, maxLength: 255
9638
9639 - `country: optional string or null`
9640
9641 Two-letter ISO 3166-1 country code, uppercase.
9642
9643 - `region: optional string or null`
9644
9645 Region or state name.
9646
9647 minLength: 1, maxLength: 255
9648
9649 - `timezone: optional string or null`
9650
9651 IANA timezone identifier, e.g. "America/Los_Angeles".
9652
9653 minLength: 1, maxLength: 255
9654
9655### Beta Managed Agents Write Tool Config
9656
9657- `BetaManagedAgentsWriteToolConfig object`
9658
9659 Configuration for the write tool.
9660
9661 - `type: "write"`
9662
9663 - `enabled: boolean`
9664
9665 - `name: "write"`
9666
9667 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9668
9669 Permission policy for tool execution.
9670
9671 - `BetaManagedAgentsAlwaysAllowPolicy object`
9672
9673 Tool calls are automatically approved without user confirmation.
9674
9675 - `type: "always_allow"`
9676
9677 - `BetaManagedAgentsAlwaysAskPolicy object`
9678
9679 Tool calls require user confirmation before execution.
9680
9681 - `type: "always_ask"`
9682
9683 - `BetaManagedAgentsAutoPolicy object`
9684
9685 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9686
9687 - `type: "auto"`
9688
9689### Beta Managed Agents Write Tool Config Params
9690
9691- `BetaManagedAgentsWriteToolConfigParams object`
9692
9693 Configuration override for the write tool.
9694
9695 - `type: optional "write"`
9696
9697 - `name: "write"`
9698
9699 Must be "write".
9700
9701 - `enabled: optional boolean or null`
9702
9703 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
9704
9705 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
9706
9707 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
9708
9709 - `BetaManagedAgentsAlwaysAllowPolicy object`
9710
9711 Tool calls are automatically approved without user confirmation.
9712
9713 - `type: "always_allow"`
9714
9715 - `BetaManagedAgentsAlwaysAskPolicy object`
9716
9717 Tool calls require user confirmation before execution.
9718
9719 - `type: "always_ask"`
9720
9721 - `BetaManagedAgentsAutoPolicy object`
9722
9723 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
9724
9725 - `type: "auto"`
9726
9727## Agents › Versions
9728
9729### List Agent Versions
9730
9731**GET** `/v1/agents/{agent_id}/versions`
9732
9733List Agent Versions
9734
9735#### Path parameters
9736
9737- `agent_id: string`
9738
9739 Agent ID to list versions for.
9740
9741#### Query parameters
9742
9743- `limit: optional number`
9744
9745 Maximum results per page. Default 20, maximum 100.
9746
9747 format: int32
9748
9749- `page: optional string`
9750
9751 Opaque pagination cursor.
9752
9753#### Headers
9754
9755- `"anthropic-beta": optional array of AnthropicBeta`
9756
9757 Optional header to specify the beta version(s) you want to use.
9758
9759 - `string`
9760
9761 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
9762
9763 - `"message-batches-2024-09-24"`
9764
9765 - `"prompt-caching-2024-07-31"`
9766
9767 - `"computer-use-2024-10-22"`
9768
9769 - `"computer-use-2025-01-24"`
9770
9771 - `"pdfs-2024-09-25"`
9772
9773 - `"token-counting-2024-11-01"`
9774
9775 - `"token-efficient-tools-2025-02-19"`
9776
9777 - `"output-128k-2025-02-19"`
9778
9779 - `"files-api-2025-04-14"`
9780
9781 - `"mcp-client-2025-04-04"`
9782
9783 - `"mcp-client-2025-11-20"`
9784
9785 - `"dev-full-thinking-2025-05-14"`
9786
9787 - `"interleaved-thinking-2025-05-14"`
9788
9789 - `"code-execution-2025-05-22"`
9790
9791 - `"extended-cache-ttl-2025-04-11"`
9792
9793 - `"context-1m-2025-08-07"`
9794
9795 - `"context-management-2025-06-27"`
9796
9797 - `"model-context-window-exceeded-2025-08-26"`
9798
9799 - `"skills-2025-10-02"`
9800
9801 - `"fast-mode-2026-02-01"`
9802
9803 - `"output-300k-2026-03-24"`
9804
9805 - `"user-profiles-2026-03-24"`
9806
9807 - `"user-profiles-2026-08-18"`
9808
9809 - `"user-profiles-2026-09-04"`
9810
9811 - `"advisor-tool-2026-03-01"`
9812
9813 - `"managed-agents-2026-04-01"`
9814
9815 - `"cache-diagnosis-2026-04-07"`
9816
9817 - `"dreaming-2026-04-21"`
9818
9819 - `"thinking-token-count-2026-05-13"`
9820
9821 - `"server-side-fallback-2026-06-01"`
9822
9823 - `"server-side-fallback-2026-07-01"`
9824
9825 - `"fallback-credit-2026-06-01"`
9826
9827 - `"fallback-credit-2026-07-01"`
9828
9829 - `"agent-memory-2026-07-22"`
9830
9831 - `"mid-conversation-tool-changes-2026-07-01"`
9832
9833 - `"compact-2026-01-12"`
9834
9835 - `"computer-use-2025-11-24"`
9836
9837 - `"mcp-tunnels-2026-06-22"`
9838
9839 - `"structured-outputs-2025-11-13"`
9840
9841 - `"task-budgets-2026-03-13"`
9842
9843 - `"thinking-display-updates-2026-08-18"`
9844
9845 - `"ce-user-management-2026-07-13"`
9846
9847 - `"mid-conversation-output-config-2026-07-01"`
9848
9849 - `"thinking-binding-controls-2026-08-01"`
9850
9851 - `"mid-conversation-system-clear-at-2026-08-21"`
9852
9853 - `"compact-2026-09-04"`
9854
9855 - `"inline-tools-2026-09-15"`
9856
9857 - `"mcp-client-2026-09-15"`
9858
9859- `"anthropic-workspace-id": optional string`
9860
9861 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
9862
9863 Only needed for credentials that can act on more than one Workspace. A credential that belongs to a specific Workspace may omit it; if sent, it must match that Workspace.
9864
9865#### Returns
9866
9867- `data: array of BetaManagedAgentsAgent`
9868
9869 Agent versions.
9870
9871 - `type: "agent"`
9872
9873 - `id: string`
9874
9875 - `archived_at: string or null`
9876
9877 When the agent was archived. Null if not archived.
9878
9879 format: date-time
9880
9881 - `created_at: string`
9882
9883 A timestamp in RFC 3339 format
9884
9885 format: date-time
9886
9887 - `description: string or null`
9888
9889 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
9890
9891 - `type: "url"`
9892
9893 - `name: string`
9894
9895 - `url: string`
9896
9897 - `metadata: map[string]`
9898
9899 - `model: BetaManagedAgentsModelConfig`
9900
9901 Model identifier and configuration.
9902
9903 - `id: BetaManagedAgentsModel`
9904
9905 The model that will power your agent.
9906
9907 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
9908
9909 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
9910
9911 The model that will power your agent.
9912
9913 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
9914
9915 - `"claude-sonnet-5-5"`
9916
9917 Efficient model for coding and agents
9918
9919 - `"claude-opus-5-5"`
9920
9921 Powerful intelligence for coding, knowledge work, and long-running agents
9922
9923 - `"claude-fable-5-1"`
9924
9925 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
9926
9927 - `"claude-sonnet-5"`
9928
9929 Efficient model for coding and agents
9930
9931 - `"claude-fable-5"`
9932
9933 Next generation of intelligence for the hardest knowledge work and coding problems
9934
9935 - `"claude-opus-5"`
9936
9937 Powerful intelligence for long-running agents and coding
9938
9939 - `"claude-opus-4-8"`
9940
9941 Powerful intelligence for long-running agents and coding
9942
9943 - `"claude-opus-4-7"`
9944
9945 Powerful intelligence for long-running agents and coding
9946
9947 - `"claude-opus-4-6"`
9948
9949 Powerful intelligence for long-running agents and coding
9950
9951 - `"claude-sonnet-4-6"`
9952
9953 Best combination of speed and intelligence
9954
9955 - `"claude-haiku-4-5"`
9956
9957 Fastest model with near-frontier intelligence
9958
9959 - `"claude-haiku-4-5-20251001"`
9960
9961 Fastest model with near-frontier intelligence
9962
9963 - `"claude-opus-4-5"`
9964
9965 Powerful intelligence for long-running agents and coding
9966
9967 - `"claude-opus-4-5-20251101"`
9968
9969 Powerful intelligence for long-running agents and coding
9970
9971 - `"claude-sonnet-4-5"`
9972
9973 High-performance model for agents and coding
9974
9975 - `"claude-sonnet-4-5-20250929"`
9976
9977 High-performance model for agents and coding
99789876
99799877 - `string`
99809878
99819879
api/beta/agents/archive Changed · +158 / -166 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
121121
122 - `"mcp-client-2026-09-15"`
123
124122- `"anthropic-workspace-id": optional string`
125123
126124 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 169
171169
172170 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
173171
174 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
172 - `"claude-sonnet-5-5"`
175173
176 The model that will power your agent.
174 Efficient model for coding and agents
177175
178 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
176 - `"claude-opus-5-5"`
179177
180 - `"claude-sonnet-5-5"`
178 Powerful intelligence for coding, knowledge work, and long-running agents
181179
182 Efficient model for coding and agents
180 - `"claude-fable-5-1"`
183181
184 - `"claude-opus-5-5"`
182 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
185183
186 Powerful intelligence for coding, knowledge work, and long-running agents
184 - `"claude-sonnet-5"`
187185
188 - `"claude-fable-5-1"`
186 Efficient model for coding and agents
189187
190 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
188 - `"claude-fable-5"`
191189
192 - `"claude-sonnet-5"`
190 Next generation of intelligence for the hardest knowledge work and coding problems
193191
194 Efficient model for coding and agents
192 - `"claude-opus-5"`
195193
196 - `"claude-fable-5"`
194 Powerful intelligence for long-running agents and coding
197195
198 Next generation of intelligence for the hardest knowledge work and coding problems
196 - `"claude-opus-4-8"`
199197
200 - `"claude-opus-5"`
198 Powerful intelligence for long-running agents and coding
201199
202 Powerful intelligence for long-running agents and coding
200 - `"claude-opus-4-7"`
203201
204 - `"claude-opus-4-8"`
202 Powerful intelligence for long-running agents and coding
205203
206 Powerful intelligence for long-running agents and coding
204 - `"claude-opus-4-6"`
207205
208 - `"claude-opus-4-7"`
206 Powerful intelligence for long-running agents and coding
209207
210 Powerful intelligence for long-running agents and coding
208 - `"claude-sonnet-4-6"`
211209
212 - `"claude-opus-4-6"`
210 Best combination of speed and intelligence
213211
214 Powerful intelligence for long-running agents and coding
212 - `"claude-haiku-4-5"`
215213
216 - `"claude-sonnet-4-6"`
214 Fastest model with near-frontier intelligence
217215
218 Best combination of speed and intelligence
216 - `"claude-haiku-4-5-20251001"`
219217
220 - `"claude-haiku-4-5"`
218 Fastest model with near-frontier intelligence
221219
222 Fastest model with near-frontier intelligence
220 - `"claude-opus-4-5"`
223221
224 - `"claude-haiku-4-5-20251001"`
222 Powerful intelligence for long-running agents and coding
225223
226 Fastest model with near-frontier intelligence
224 - `"claude-opus-4-5-20251101"`
227225
228 - `"claude-opus-4-5"`
226 Powerful intelligence for long-running agents and coding
229227
230 Powerful intelligence for long-running agents and coding
228 - `"claude-sonnet-4-5"`
231229
232 - `"claude-opus-4-5-20251101"`
230 High-performance model for agents and coding
233231
234 Powerful intelligence for long-running agents and coding
232 - `"claude-sonnet-4-5-20250929"`
235233
236 - `"claude-sonnet-4-5"`
237
238 High-performance model for agents and coding
239
240 - `"claude-sonnet-4-5-20250929"`
241
242 High-performance model for agents and coding
234 High-performance model for agents and coding
243235
244236 - `string`
245237
api/beta/agents/create Changed · +834 / -852 lines
from line 17
1717
1818 - `string`
1919
20 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
20 - `"message-batches-2024-09-24"`
2121
22 - `"message-batches-2024-09-24"`
22 - `"prompt-caching-2024-07-31"`
2323
24 - `"prompt-caching-2024-07-31"`
24 - `"computer-use-2024-10-22"`
2525
26 - `"computer-use-2024-10-22"`
26 - `"computer-use-2025-01-24"`
2727
28 - `"computer-use-2025-01-24"`
28 - `"pdfs-2024-09-25"`
2929
30 - `"pdfs-2024-09-25"`
30 - `"token-counting-2024-11-01"`
3131
32 - `"token-counting-2024-11-01"`
32 - `"token-efficient-tools-2025-02-19"`
3333
34 - `"token-efficient-tools-2025-02-19"`
34 - `"output-128k-2025-02-19"`
3535
36 - `"output-128k-2025-02-19"`
36 - `"files-api-2025-04-14"`
3737
38 - `"files-api-2025-04-14"`
38 - `"mcp-client-2025-04-04"`
3939
40 - `"mcp-client-2025-04-04"`
40 - `"mcp-client-2025-11-20"`
4141
42 - `"mcp-client-2025-11-20"`
42 - `"dev-full-thinking-2025-05-14"`
4343
44 - `"dev-full-thinking-2025-05-14"`
44 - `"interleaved-thinking-2025-05-14"`
4545
46 - `"interleaved-thinking-2025-05-14"`
46 - `"code-execution-2025-05-22"`
4747
48 - `"code-execution-2025-05-22"`
48 - `"extended-cache-ttl-2025-04-11"`
4949
50 - `"extended-cache-ttl-2025-04-11"`
50 - `"context-1m-2025-08-07"`
5151
52 - `"context-1m-2025-08-07"`
52 - `"context-management-2025-06-27"`
5353
54 - `"context-management-2025-06-27"`
54 - `"model-context-window-exceeded-2025-08-26"`
5555
56 - `"model-context-window-exceeded-2025-08-26"`
56 - `"skills-2025-10-02"`
5757
58 - `"skills-2025-10-02"`
58 - `"fast-mode-2026-02-01"`
5959
60 - `"fast-mode-2026-02-01"`
60 - `"output-300k-2026-03-24"`
6161
62 - `"output-300k-2026-03-24"`
62 - `"user-profiles-2026-03-24"`
6363
64 - `"user-profiles-2026-03-24"`
64 - `"user-profiles-2026-08-18"`
6565
66 - `"user-profiles-2026-08-18"`
66 - `"user-profiles-2026-09-04"`
6767
68 - `"user-profiles-2026-09-04"`
68 - `"advisor-tool-2026-03-01"`
6969
70 - `"advisor-tool-2026-03-01"`
70 - `"managed-agents-2026-04-01"`
7171
72 - `"managed-agents-2026-04-01"`
72 - `"cache-diagnosis-2026-04-07"`
7373
74 - `"cache-diagnosis-2026-04-07"`
74 - `"dreaming-2026-04-21"`
7575
76 - `"dreaming-2026-04-21"`
76 - `"thinking-token-count-2026-05-13"`
7777
78 - `"thinking-token-count-2026-05-13"`
78 - `"server-side-fallback-2026-06-01"`
7979
80 - `"server-side-fallback-2026-06-01"`
80 - `"server-side-fallback-2026-07-01"`
8181
82 - `"server-side-fallback-2026-07-01"`
82 - `"fallback-credit-2026-06-01"`
8383
84 - `"fallback-credit-2026-06-01"`
84 - `"fallback-credit-2026-07-01"`
8585
86 - `"fallback-credit-2026-07-01"`
86 - `"agent-memory-2026-07-22"`
8787
88 - `"agent-memory-2026-07-22"`
88 - `"mid-conversation-tool-changes-2026-07-01"`
8989
90 - `"mid-conversation-tool-changes-2026-07-01"`
90 - `"compact-2026-01-12"`
9191
92 - `"compact-2026-01-12"`
92 - `"computer-use-2025-11-24"`
9393
94 - `"computer-use-2025-11-24"`
94 - `"mcp-tunnels-2026-06-22"`
9595
96 - `"mcp-tunnels-2026-06-22"`
96 - `"structured-outputs-2025-11-13"`
9797
98 - `"structured-outputs-2025-11-13"`
98 - `"task-budgets-2026-03-13"`
9999
100 - `"task-budgets-2026-03-13"`
100 - `"thinking-display-updates-2026-08-18"`
101101
102 - `"thinking-display-updates-2026-08-18"`
102 - `"ce-user-management-2026-07-13"`
103103
104 - `"ce-user-management-2026-07-13"`
104 - `"mid-conversation-output-config-2026-07-01"`
105105
106 - `"mid-conversation-output-config-2026-07-01"`
106 - `"thinking-binding-controls-2026-08-01"`
107107
108 - `"thinking-binding-controls-2026-08-01"`
108 - `"mid-conversation-system-clear-at-2026-08-21"`
109109
110 - `"mid-conversation-system-clear-at-2026-08-21"`
110 - `"compact-2026-09-04"`
111111
112 - `"compact-2026-09-04"`
112 - `"inline-tools-2026-09-15"`
113113
114 - `"inline-tools-2026-09-15"`
114 - `"mcp-client-2026-09-15"`
115115
116 - `"mcp-client-2026-09-15"`
117
118116- `"anthropic-workspace-id": optional string`
119117
120118 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 131
133131
134132 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
135133
136 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
134 - `"claude-sonnet-5-5"`
137135
138 The model that will power your agent.
136 Efficient model for coding and agents
139137
140 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
138 - `"claude-opus-5-5"`
141139
142 - `"claude-sonnet-5-5"`
140 Powerful intelligence for coding, knowledge work, and long-running agents
143141
144 Efficient model for coding and agents
142 - `"claude-fable-5-1"`
145143
146 - `"claude-opus-5-5"`
144 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
147145
148 Powerful intelligence for coding, knowledge work, and long-running agents
146 - `"claude-sonnet-5"`
149147
150 - `"claude-fable-5-1"`
148 Efficient model for coding and agents
151149
152 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
150 - `"claude-fable-5"`
153151
154 - `"claude-sonnet-5"`
152 Next generation of intelligence for the hardest knowledge work and coding problems
155153
156 Efficient model for coding and agents
154 - `"claude-opus-5"`
157155
158 - `"claude-fable-5"`
156 Powerful intelligence for long-running agents and coding
159157
160 Next generation of intelligence for the hardest knowledge work and coding problems
158 - `"claude-opus-4-8"`
161159
162 - `"claude-opus-5"`
160 Powerful intelligence for long-running agents and coding
163161
164 Powerful intelligence for long-running agents and coding
162 - `"claude-opus-4-7"`
165163
166 - `"claude-opus-4-8"`
164 Powerful intelligence for long-running agents and coding
167165
168 Powerful intelligence for long-running agents and coding
166 - `"claude-opus-4-6"`
169167
170 - `"claude-opus-4-7"`
168 Powerful intelligence for long-running agents and coding
171169
172 Powerful intelligence for long-running agents and coding
170 - `"claude-sonnet-4-6"`
173171
174 - `"claude-opus-4-6"`
172 Best combination of speed and intelligence
175173
176 Powerful intelligence for long-running agents and coding
174 - `"claude-haiku-4-5"`
177175
178 - `"claude-sonnet-4-6"`
176 Fastest model with near-frontier intelligence
179177
180 Best combination of speed and intelligence
178 - `"claude-haiku-4-5-20251001"`
181179
182 - `"claude-haiku-4-5"`
180 Fastest model with near-frontier intelligence
183181
184 Fastest model with near-frontier intelligence
182 - `"claude-opus-4-5"`
185183
186 - `"claude-haiku-4-5-20251001"`
184 Powerful intelligence for long-running agents and coding
187185
188 Fastest model with near-frontier intelligence
186 - `"claude-opus-4-5-20251101"`
189187
190 - `"claude-opus-4-5"`
188 Powerful intelligence for long-running agents and coding
191189
192 Powerful intelligence for long-running agents and coding
190 - `"claude-sonnet-4-5"`
193191
194 - `"claude-opus-4-5-20251101"`
192 High-performance model for agents and coding
195193
196 Powerful intelligence for long-running agents and coding
194 - `"claude-sonnet-4-5-20250929"`
197195
198 - `"claude-sonnet-4-5"`
196 High-performance model for agents and coding
199197
200 High-performance model for agents and coding
201
202 - `"claude-sonnet-4-5-20250929"`
203
204 High-performance model for agents and coding
205
206198 - `string`
207199
208200 - `BetaManagedAgentsModelConfigParams object`
from line 211
219211
220212 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
221213
222 - `BetaManagedAgentsEffortLevel = "low" or "medium" or "high" or 2 more`
214 - `"low"`
223215
224 How hard Claude works on each turn. Higher levels favor reasoning depth over latency. Not all models accept every level; invalid combinations are rejected at create time.
216 Low effort. Favors latency over reasoning depth.
225217
226 - `"low"`
218 - `"medium"`
227219
228 Low effort. Favors latency over reasoning depth.
220 Medium effort. Balances latency and reasoning depth.
229221
230 - `"medium"`
222 - `"high"`
231223
232 Medium effort. Balances latency and reasoning depth.
224 High effort. Favors reasoning depth.
233225
234 - `"high"`
226 - `"xhigh"`
235227
236 High effort. Favors reasoning depth.
228 Extra-high effort. Not all models accept this level.
237229
238 - `"xhigh"`
230 - `"max"`
239231
240 Extra-high effort. Not all models accept this level.
232 Maximum effort. Favors reasoning depth over latency.
241233
242 - `"max"`
243
244 Maximum effort. Favors reasoning depth over latency.
245
246234 - `BetaManagedAgentsEffortLow object`
247235
248236 Low effort. Favors latency over reasoning depth.
from line 875
887875
888876 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
889877
890 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
878 - `"claude-sonnet-5-5"`
891879
892 The model that will power your agent.
880 Efficient model for coding and agents
893881
894 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
882 - `"claude-opus-5-5"`
895883
896 - `"claude-sonnet-5-5"`
884 Powerful intelligence for coding, knowledge work, and long-running agents
897885
898 Efficient model for coding and agents
886 - `"claude-fable-5-1"`
899887
900 - `"claude-opus-5-5"`
888 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
901889
902 Powerful intelligence for coding, knowledge work, and long-running agents
890 - `"claude-sonnet-5"`
903891
904 - `"claude-fable-5-1"`
892 Efficient model for coding and agents
905893
906 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
894 - `"claude-fable-5"`
907895
908 - `"claude-sonnet-5"`
896 Next generation of intelligence for the hardest knowledge work and coding problems
909897
910 Efficient model for coding and agents
898 - `"claude-opus-5"`
911899
912 - `"claude-fable-5"`
900 Powerful intelligence for long-running agents and coding
913901
914 Next generation of intelligence for the hardest knowledge work and coding problems
902 - `"claude-opus-4-8"`
915903
916 - `"claude-opus-5"`
904 Powerful intelligence for long-running agents and coding
917905
918 Powerful intelligence for long-running agents and coding
906 - `"claude-opus-4-7"`
919907
920 - `"claude-opus-4-8"`
908 Powerful intelligence for long-running agents and coding
921909
922 Powerful intelligence for long-running agents and coding
910 - `"claude-opus-4-6"`
923911
924 - `"claude-opus-4-7"`
912 Powerful intelligence for long-running agents and coding
925913
926 Powerful intelligence for long-running agents and coding
914 - `"claude-sonnet-4-6"`
927915
928 - `"claude-opus-4-6"`
916 Best combination of speed and intelligence
929917
930 Powerful intelligence for long-running agents and coding
918 - `"claude-haiku-4-5"`
931919
932 - `"claude-sonnet-4-6"`
920 Fastest model with near-frontier intelligence
933921
934 Best combination of speed and intelligence
922 - `"claude-haiku-4-5-20251001"`
935923
936 - `"claude-haiku-4-5"`
924 Fastest model with near-frontier intelligence
937925
938 Fastest model with near-frontier intelligence
926 - `"claude-opus-4-5"`
939927
940 - `"claude-haiku-4-5-20251001"`
928 Powerful intelligence for long-running agents and coding
941929
942 Fastest model with near-frontier intelligence
930 - `"claude-opus-4-5-20251101"`
943931
944 - `"claude-opus-4-5"`
932 Powerful intelligence for long-running agents and coding
945933
946 Powerful intelligence for long-running agents and coding
934 - `"claude-sonnet-4-5"`
947935
948 - `"claude-opus-4-5-20251101"`
936 High-performance model for agents and coding
949937
950 Powerful intelligence for long-running agents and coding
938 - `"claude-sonnet-4-5-20250929"`
951939
952 - `"claude-sonnet-4-5"`
953
954 High-performance model for agents and coding
955
956 - `"claude-sonnet-4-5-20250929"`
957
958 High-performance model for agents and coding
940 High-performance model for agents and coding
959941
960942 - `string`
961943
api/beta/agents/list Changed · +158 / -166 lines
from line 45
4545
4646 - `string`
4747
48 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
48 - `"message-batches-2024-09-24"`
4949
50 - `"message-batches-2024-09-24"`
50 - `"prompt-caching-2024-07-31"`
5151
52 - `"prompt-caching-2024-07-31"`
52 - `"computer-use-2024-10-22"`
5353
54 - `"computer-use-2024-10-22"`
54 - `"computer-use-2025-01-24"`
5555
56 - `"computer-use-2025-01-24"`
56 - `"pdfs-2024-09-25"`
5757
58 - `"pdfs-2024-09-25"`
58 - `"token-counting-2024-11-01"`
5959
60 - `"token-counting-2024-11-01"`
60 - `"token-efficient-tools-2025-02-19"`
6161
62 - `"token-efficient-tools-2025-02-19"`
62 - `"output-128k-2025-02-19"`
6363
64 - `"output-128k-2025-02-19"`
64 - `"files-api-2025-04-14"`
6565
66 - `"files-api-2025-04-14"`
66 - `"mcp-client-2025-04-04"`
6767
68 - `"mcp-client-2025-04-04"`
68 - `"mcp-client-2025-11-20"`
6969
70 - `"mcp-client-2025-11-20"`
70 - `"dev-full-thinking-2025-05-14"`
7171
72 - `"dev-full-thinking-2025-05-14"`
72 - `"interleaved-thinking-2025-05-14"`
7373
74 - `"interleaved-thinking-2025-05-14"`
74 - `"code-execution-2025-05-22"`
7575
76 - `"code-execution-2025-05-22"`
76 - `"extended-cache-ttl-2025-04-11"`
7777
78 - `"extended-cache-ttl-2025-04-11"`
78 - `"context-1m-2025-08-07"`
7979
80 - `"context-1m-2025-08-07"`
80 - `"context-management-2025-06-27"`
8181
82 - `"context-management-2025-06-27"`
82 - `"model-context-window-exceeded-2025-08-26"`
8383
84 - `"model-context-window-exceeded-2025-08-26"`
84 - `"skills-2025-10-02"`
8585
86 - `"skills-2025-10-02"`
86 - `"fast-mode-2026-02-01"`
8787
88 - `"fast-mode-2026-02-01"`
88 - `"output-300k-2026-03-24"`
8989
90 - `"output-300k-2026-03-24"`
90 - `"user-profiles-2026-03-24"`
9191
92 - `"user-profiles-2026-03-24"`
92 - `"user-profiles-2026-08-18"`
9393
94 - `"user-profiles-2026-08-18"`
94 - `"user-profiles-2026-09-04"`
9595
96 - `"user-profiles-2026-09-04"`
96 - `"advisor-tool-2026-03-01"`
9797
98 - `"advisor-tool-2026-03-01"`
98 - `"managed-agents-2026-04-01"`
9999
100 - `"managed-agents-2026-04-01"`
100 - `"cache-diagnosis-2026-04-07"`
101101
102 - `"cache-diagnosis-2026-04-07"`
102 - `"dreaming-2026-04-21"`
103103
104 - `"dreaming-2026-04-21"`
104 - `"thinking-token-count-2026-05-13"`
105105
106 - `"thinking-token-count-2026-05-13"`
106 - `"server-side-fallback-2026-06-01"`
107107
108 - `"server-side-fallback-2026-06-01"`
108 - `"server-side-fallback-2026-07-01"`
109109
110 - `"server-side-fallback-2026-07-01"`
110 - `"fallback-credit-2026-06-01"`
111111
112 - `"fallback-credit-2026-06-01"`
112 - `"fallback-credit-2026-07-01"`
113113
114 - `"fallback-credit-2026-07-01"`
114 - `"agent-memory-2026-07-22"`
115115
116 - `"agent-memory-2026-07-22"`
116 - `"mid-conversation-tool-changes-2026-07-01"`
117117
118 - `"mid-conversation-tool-changes-2026-07-01"`
118 - `"compact-2026-01-12"`
119119
120 - `"compact-2026-01-12"`
120 - `"computer-use-2025-11-24"`
121121
122 - `"computer-use-2025-11-24"`
122 - `"mcp-tunnels-2026-06-22"`
123123
124 - `"mcp-tunnels-2026-06-22"`
124 - `"structured-outputs-2025-11-13"`
125125
126 - `"structured-outputs-2025-11-13"`
126 - `"task-budgets-2026-03-13"`
127127
128 - `"task-budgets-2026-03-13"`
128 - `"thinking-display-updates-2026-08-18"`
129129
130 - `"thinking-display-updates-2026-08-18"`
130 - `"ce-user-management-2026-07-13"`
131131
132 - `"ce-user-management-2026-07-13"`
132 - `"mid-conversation-output-config-2026-07-01"`
133133
134 - `"mid-conversation-output-config-2026-07-01"`
134 - `"thinking-binding-controls-2026-08-01"`
135135
136 - `"thinking-binding-controls-2026-08-01"`
136 - `"mid-conversation-system-clear-at-2026-08-21"`
137137
138 - `"mid-conversation-system-clear-at-2026-08-21"`
138 - `"compact-2026-09-04"`
139139
140 - `"compact-2026-09-04"`
140 - `"inline-tools-2026-09-15"`
141141
142 - `"inline-tools-2026-09-15"`
142 - `"mcp-client-2026-09-15"`
143143
144 - `"mcp-client-2026-09-15"`
145
146144- `"anthropic-workspace-id": optional string`
147145
148146 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 191
193191
194192 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
195193
196 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
194 - `"claude-sonnet-5-5"`
197195
198 The model that will power your agent.
196 Efficient model for coding and agents
199197
200 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
198 - `"claude-opus-5-5"`
201199
202 - `"claude-sonnet-5-5"`
200 Powerful intelligence for coding, knowledge work, and long-running agents
203201
204 Efficient model for coding and agents
202 - `"claude-fable-5-1"`
205203
206 - `"claude-opus-5-5"`
204 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
207205
208 Powerful intelligence for coding, knowledge work, and long-running agents
206 - `"claude-sonnet-5"`
209207
210 - `"claude-fable-5-1"`
208 Efficient model for coding and agents
211209
212 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
210 - `"claude-fable-5"`
213211
214 - `"claude-sonnet-5"`
212 Next generation of intelligence for the hardest knowledge work and coding problems
215213
216 Efficient model for coding and agents
214 - `"claude-opus-5"`
217215
218 - `"claude-fable-5"`
216 Powerful intelligence for long-running agents and coding
219217
220 Next generation of intelligence for the hardest knowledge work and coding problems
218 - `"claude-opus-4-8"`
221219
222 - `"claude-opus-5"`
220 Powerful intelligence for long-running agents and coding
223221
224 Powerful intelligence for long-running agents and coding
222 - `"claude-opus-4-7"`
225223
226 - `"claude-opus-4-8"`
224 Powerful intelligence for long-running agents and coding
227225
228 Powerful intelligence for long-running agents and coding
226 - `"claude-opus-4-6"`
229227
230 - `"claude-opus-4-7"`
228 Powerful intelligence for long-running agents and coding
231229
232 Powerful intelligence for long-running agents and coding
230 - `"claude-sonnet-4-6"`
233231
234 - `"claude-opus-4-6"`
232 Best combination of speed and intelligence
235233
236 Powerful intelligence for long-running agents and coding
234 - `"claude-haiku-4-5"`
237235
238 - `"claude-sonnet-4-6"`
236 Fastest model with near-frontier intelligence
239237
240 Best combination of speed and intelligence
238 - `"claude-haiku-4-5-20251001"`
241239
242 - `"claude-haiku-4-5"`
240 Fastest model with near-frontier intelligence
243241
244 Fastest model with near-frontier intelligence
242 - `"claude-opus-4-5"`
245243
246 - `"claude-haiku-4-5-20251001"`
244 Powerful intelligence for long-running agents and coding
247245
248 Fastest model with near-frontier intelligence
246 - `"claude-opus-4-5-20251101"`
249247
250 - `"claude-opus-4-5"`
248 Powerful intelligence for long-running agents and coding
251249
252 Powerful intelligence for long-running agents and coding
250 - `"claude-sonnet-4-5"`
253251
254 - `"claude-opus-4-5-20251101"`
252 High-performance model for agents and coding
255253
256 Powerful intelligence for long-running agents and coding
254 - `"claude-sonnet-4-5-20250929"`
257255
258 - `"claude-sonnet-4-5"`
259
260 High-performance model for agents and coding
261
262 - `"claude-sonnet-4-5-20250929"`
263
264 High-performance model for agents and coding
256 High-performance model for agents and coding
265257
266258 - `string`
267259
api/beta/agents/retrieve Changed · +158 / -166 lines
from line 31
3131
3232 - `string`
3333
34 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
34 - `"message-batches-2024-09-24"`
3535
36 - `"message-batches-2024-09-24"`
36 - `"prompt-caching-2024-07-31"`
3737
38 - `"prompt-caching-2024-07-31"`
38 - `"computer-use-2024-10-22"`
3939
40 - `"computer-use-2024-10-22"`
40 - `"computer-use-2025-01-24"`
4141
42 - `"computer-use-2025-01-24"`
42 - `"pdfs-2024-09-25"`
4343
44 - `"pdfs-2024-09-25"`
44 - `"token-counting-2024-11-01"`
4545
46 - `"token-counting-2024-11-01"`
46 - `"token-efficient-tools-2025-02-19"`
4747
48 - `"token-efficient-tools-2025-02-19"`
48 - `"output-128k-2025-02-19"`
4949
50 - `"output-128k-2025-02-19"`
50 - `"files-api-2025-04-14"`
5151
52 - `"files-api-2025-04-14"`
52 - `"mcp-client-2025-04-04"`
5353
54 - `"mcp-client-2025-04-04"`
54 - `"mcp-client-2025-11-20"`
5555
56 - `"mcp-client-2025-11-20"`
56 - `"dev-full-thinking-2025-05-14"`
5757
58 - `"dev-full-thinking-2025-05-14"`
58 - `"interleaved-thinking-2025-05-14"`
5959
60 - `"interleaved-thinking-2025-05-14"`
60 - `"code-execution-2025-05-22"`
6161
62 - `"code-execution-2025-05-22"`
62 - `"extended-cache-ttl-2025-04-11"`
6363
64 - `"extended-cache-ttl-2025-04-11"`
64 - `"context-1m-2025-08-07"`
6565
66 - `"context-1m-2025-08-07"`
66 - `"context-management-2025-06-27"`
6767
68 - `"context-management-2025-06-27"`
68 - `"model-context-window-exceeded-2025-08-26"`
6969
70 - `"model-context-window-exceeded-2025-08-26"`
70 - `"skills-2025-10-02"`
7171
72 - `"skills-2025-10-02"`
72 - `"fast-mode-2026-02-01"`
7373
74 - `"fast-mode-2026-02-01"`
74 - `"output-300k-2026-03-24"`
7575
76 - `"output-300k-2026-03-24"`
76 - `"user-profiles-2026-03-24"`
7777
78 - `"user-profiles-2026-03-24"`
78 - `"user-profiles-2026-08-18"`
7979
80 - `"user-profiles-2026-08-18"`
80 - `"user-profiles-2026-09-04"`
8181
82 - `"user-profiles-2026-09-04"`
82 - `"advisor-tool-2026-03-01"`
8383
84 - `"advisor-tool-2026-03-01"`
84 - `"managed-agents-2026-04-01"`
8585
86 - `"managed-agents-2026-04-01"`
86 - `"cache-diagnosis-2026-04-07"`
8787
88 - `"cache-diagnosis-2026-04-07"`
88 - `"dreaming-2026-04-21"`
8989
90 - `"dreaming-2026-04-21"`
90 - `"thinking-token-count-2026-05-13"`
9191
92 - `"thinking-token-count-2026-05-13"`
92 - `"server-side-fallback-2026-06-01"`
9393
94 - `"server-side-fallback-2026-06-01"`
94 - `"server-side-fallback-2026-07-01"`
9595
96 - `"server-side-fallback-2026-07-01"`
96 - `"fallback-credit-2026-06-01"`
9797
98 - `"fallback-credit-2026-06-01"`
98 - `"fallback-credit-2026-07-01"`
9999
100 - `"fallback-credit-2026-07-01"`
100 - `"agent-memory-2026-07-22"`
101101
102 - `"agent-memory-2026-07-22"`
102 - `"mid-conversation-tool-changes-2026-07-01"`
103103
104 - `"mid-conversation-tool-changes-2026-07-01"`
104 - `"compact-2026-01-12"`
105105
106 - `"compact-2026-01-12"`
106 - `"computer-use-2025-11-24"`
107107
108 - `"computer-use-2025-11-24"`
108 - `"mcp-tunnels-2026-06-22"`
109109
110 - `"mcp-tunnels-2026-06-22"`
110 - `"structured-outputs-2025-11-13"`
111111
112 - `"structured-outputs-2025-11-13"`
112 - `"task-budgets-2026-03-13"`
113113
114 - `"task-budgets-2026-03-13"`
114 - `"thinking-display-updates-2026-08-18"`
115115
116 - `"thinking-display-updates-2026-08-18"`
116 - `"ce-user-management-2026-07-13"`
117117
118 - `"ce-user-management-2026-07-13"`
118 - `"mid-conversation-output-config-2026-07-01"`
119119
120 - `"mid-conversation-output-config-2026-07-01"`
120 - `"thinking-binding-controls-2026-08-01"`
121121
122 - `"thinking-binding-controls-2026-08-01"`
122 - `"mid-conversation-system-clear-at-2026-08-21"`
123123
124 - `"mid-conversation-system-clear-at-2026-08-21"`
124 - `"compact-2026-09-04"`
125125
126 - `"compact-2026-09-04"`
126 - `"inline-tools-2026-09-15"`
127127
128 - `"inline-tools-2026-09-15"`
128 - `"mcp-client-2026-09-15"`
129129
130 - `"mcp-client-2026-09-15"`
131
132130- `"anthropic-workspace-id": optional string`
133131
134132 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 177
179177
180178 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
181179
182 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
180 - `"claude-sonnet-5-5"`
183181
184 The model that will power your agent.
182 Efficient model for coding and agents
185183
186 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
184 - `"claude-opus-5-5"`
187185
188 - `"claude-sonnet-5-5"`
186 Powerful intelligence for coding, knowledge work, and long-running agents
189187
190 Efficient model for coding and agents
188 - `"claude-fable-5-1"`
191189
192 - `"claude-opus-5-5"`
190 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
193191
194 Powerful intelligence for coding, knowledge work, and long-running agents
192 - `"claude-sonnet-5"`
195193
196 - `"claude-fable-5-1"`
194 Efficient model for coding and agents
197195
198 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
196 - `"claude-fable-5"`
199197
200 - `"claude-sonnet-5"`
198 Next generation of intelligence for the hardest knowledge work and coding problems
201199
202 Efficient model for coding and agents
200 - `"claude-opus-5"`
203201
204 - `"claude-fable-5"`
202 Powerful intelligence for long-running agents and coding
205203
206 Next generation of intelligence for the hardest knowledge work and coding problems
204 - `"claude-opus-4-8"`
207205
208 - `"claude-opus-5"`
206 Powerful intelligence for long-running agents and coding
209207
210 Powerful intelligence for long-running agents and coding
208 - `"claude-opus-4-7"`
211209
212 - `"claude-opus-4-8"`
210 Powerful intelligence for long-running agents and coding
213211
214 Powerful intelligence for long-running agents and coding
212 - `"claude-opus-4-6"`
215213
216 - `"claude-opus-4-7"`
214 Powerful intelligence for long-running agents and coding
217215
218 Powerful intelligence for long-running agents and coding
216 - `"claude-sonnet-4-6"`
219217
220 - `"claude-opus-4-6"`
218 Best combination of speed and intelligence
221219
222 Powerful intelligence for long-running agents and coding
220 - `"claude-haiku-4-5"`
223221
224 - `"claude-sonnet-4-6"`
222 Fastest model with near-frontier intelligence
225223
226 Best combination of speed and intelligence
224 - `"claude-haiku-4-5-20251001"`
227225
228 - `"claude-haiku-4-5"`
226 Fastest model with near-frontier intelligence
229227
230 Fastest model with near-frontier intelligence
228 - `"claude-opus-4-5"`
231229
232 - `"claude-haiku-4-5-20251001"`
230 Powerful intelligence for long-running agents and coding
233231
234 Fastest model with near-frontier intelligence
232 - `"claude-opus-4-5-20251101"`
235233
236 - `"claude-opus-4-5"`
234 Powerful intelligence for long-running agents and coding
237235
238 Powerful intelligence for long-running agents and coding
236 - `"claude-sonnet-4-5"`
239237
240 - `"claude-opus-4-5-20251101"`
238 High-performance model for agents and coding
241239
242 Powerful intelligence for long-running agents and coding
240 - `"claude-sonnet-4-5-20250929"`
243241
244 - `"claude-sonnet-4-5"`
245
246 High-performance model for agents and coding
247
248 - `"claude-sonnet-4-5-20250929"`
249
250 High-performance model for agents and coding
242 High-performance model for agents and coding
251243
252244 - `string`
253245
api/beta/agents/update Changed · +812 / -830 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
121121
122 - `"mcp-client-2026-09-15"`
123
124122- `"anthropic-workspace-id": optional string`
125123
126124 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 165
167165
168166 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
169167
170 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
168 - `"claude-sonnet-5-5"`
171169
172 The model that will power your agent.
170 Efficient model for coding and agents
173171
174 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
172 - `"claude-opus-5-5"`
175173
176 - `"claude-sonnet-5-5"`
174 Powerful intelligence for coding, knowledge work, and long-running agents
177175
178 Efficient model for coding and agents
176 - `"claude-fable-5-1"`
179177
180 - `"claude-opus-5-5"`
178 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
181179
182 Powerful intelligence for coding, knowledge work, and long-running agents
180 - `"claude-sonnet-5"`
183181
184 - `"claude-fable-5-1"`
182 Efficient model for coding and agents
185183
186 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
184 - `"claude-fable-5"`
187185
188 - `"claude-sonnet-5"`
186 Next generation of intelligence for the hardest knowledge work and coding problems
189187
190 Efficient model for coding and agents
188 - `"claude-opus-5"`
191189
192 - `"claude-fable-5"`
190 Powerful intelligence for long-running agents and coding
193191
194 Next generation of intelligence for the hardest knowledge work and coding problems
192 - `"claude-opus-4-8"`
195193
196 - `"claude-opus-5"`
194 Powerful intelligence for long-running agents and coding
197195
198 Powerful intelligence for long-running agents and coding
196 - `"claude-opus-4-7"`
199197
200 - `"claude-opus-4-8"`
198 Powerful intelligence for long-running agents and coding
201199
202 Powerful intelligence for long-running agents and coding
200 - `"claude-opus-4-6"`
203201
204 - `"claude-opus-4-7"`
202 Powerful intelligence for long-running agents and coding
205203
206 Powerful intelligence for long-running agents and coding
204 - `"claude-sonnet-4-6"`
207205
208 - `"claude-opus-4-6"`
206 Best combination of speed and intelligence
209207
210 Powerful intelligence for long-running agents and coding
208 - `"claude-haiku-4-5"`
211209
212 - `"claude-sonnet-4-6"`
210 Fastest model with near-frontier intelligence
213211
214 Best combination of speed and intelligence
212 - `"claude-haiku-4-5-20251001"`
215213
216 - `"claude-haiku-4-5"`
214 Fastest model with near-frontier intelligence
217215
218 Fastest model with near-frontier intelligence
216 - `"claude-opus-4-5"`
219217
220 - `"claude-haiku-4-5-20251001"`
218 Powerful intelligence for long-running agents and coding
221219
222 Fastest model with near-frontier intelligence
220 - `"claude-opus-4-5-20251101"`
223221
224 - `"claude-opus-4-5"`
222 Powerful intelligence for long-running agents and coding
225223
226 Powerful intelligence for long-running agents and coding
224 - `"claude-sonnet-4-5"`
227225
228 - `"claude-opus-4-5-20251101"`
226 High-performance model for agents and coding
229227
230 Powerful intelligence for long-running agents and coding
228 - `"claude-sonnet-4-5-20250929"`
231229
232 - `"claude-sonnet-4-5"`
230 High-performance model for agents and coding
233231
234 High-performance model for agents and coding
235
236 - `"claude-sonnet-4-5-20250929"`
237
238 High-performance model for agents and coding
239
240232 - `string`
241233
242234 - `BetaManagedAgentsModelConfigParams object`
from line 245
253245
254246 How hard Claude works on each inference call. Accepts a bare level string (`"high"`) or `{"type": "high"}`. On create, omitting it resolves the per-model default; on update, omitting it leaves the stored value unchanged.
255247
256 - `BetaManagedAgentsEffortLevel = "low" or "medium" or "high" or 2 more`
248 - `"low"`
257249
258 How hard Claude works on each turn. Higher levels favor reasoning depth over latency. Not all models accept every level; invalid combinations are rejected at create time.
250 Low effort. Favors latency over reasoning depth.
259251
260 - `"low"`
252 - `"medium"`
261253
262 Low effort. Favors latency over reasoning depth.
254 Medium effort. Balances latency and reasoning depth.
263255
264 - `"medium"`
256 - `"high"`
265257
266 Medium effort. Balances latency and reasoning depth.
258 High effort. Favors reasoning depth.
267259
268 - `"high"`
260 - `"xhigh"`
269261
270 High effort. Favors reasoning depth.
262 Extra-high effort. Not all models accept this level.
271263
272 - `"xhigh"`
264 - `"max"`
273265
274 Extra-high effort. Not all models accept this level.
266 Maximum effort. Favors reasoning depth over latency.
275267
276 - `"max"`
277
278 Maximum effort. Favors reasoning depth over latency.
279
280268 - `BetaManagedAgentsEffortLow object`
281269
282270 Low effort. Favors latency over reasoning depth.
from line 887
899887
900888 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
901889
902 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
890 - `"claude-sonnet-5-5"`
903891
904 The model that will power your agent.
892 Efficient model for coding and agents
905893
906 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
894 - `"claude-opus-5-5"`
907895
908 - `"claude-sonnet-5-5"`
896 Powerful intelligence for coding, knowledge work, and long-running agents
909897
910 Efficient model for coding and agents
898 - `"claude-fable-5-1"`
911899
912 - `"claude-opus-5-5"`
900 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
913901
914 Powerful intelligence for coding, knowledge work, and long-running agents
902 - `"claude-sonnet-5"`
915903
916 - `"claude-fable-5-1"`
904 Efficient model for coding and agents
917905
918 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
906 - `"claude-fable-5"`
919907
920 - `"claude-sonnet-5"`
908 Next generation of intelligence for the hardest knowledge work and coding problems
921909
922 Efficient model for coding and agents
910 - `"claude-opus-5"`
923911
924 - `"claude-fable-5"`
912 Powerful intelligence for long-running agents and coding
925913
926 Next generation of intelligence for the hardest knowledge work and coding problems
914 - `"claude-opus-4-8"`
927915
928 - `"claude-opus-5"`
916 Powerful intelligence for long-running agents and coding
929917
930 Powerful intelligence for long-running agents and coding
918 - `"claude-opus-4-7"`
931919
932 - `"claude-opus-4-8"`
920 Powerful intelligence for long-running agents and coding
933921
934 Powerful intelligence for long-running agents and coding
922 - `"claude-opus-4-6"`
935923
936 - `"claude-opus-4-7"`
924 Powerful intelligence for long-running agents and coding
937925
938 Powerful intelligence for long-running agents and coding
926 - `"claude-sonnet-4-6"`
939927
940 - `"claude-opus-4-6"`
928 Best combination of speed and intelligence
941929
942 Powerful intelligence for long-running agents and coding
930 - `"claude-haiku-4-5"`
943931
944 - `"claude-sonnet-4-6"`
932 Fastest model with near-frontier intelligence
945933
946 Best combination of speed and intelligence
934 - `"claude-haiku-4-5-20251001"`
947935
948 - `"claude-haiku-4-5"`
936 Fastest model with near-frontier intelligence
949937
950 Fastest model with near-frontier intelligence
938 - `"claude-opus-4-5"`
951939
952 - `"claude-haiku-4-5-20251001"`
940 Powerful intelligence for long-running agents and coding
953941
954 Fastest model with near-frontier intelligence
942 - `"claude-opus-4-5-20251101"`
955943
956 - `"claude-opus-4-5"`
944 Powerful intelligence for long-running agents and coding
957945
958 Powerful intelligence for long-running agents and coding
946 - `"claude-sonnet-4-5"`
959947
960 - `"claude-opus-4-5-20251101"`
948 High-performance model for agents and coding
961949
962 Powerful intelligence for long-running agents and coding
950 - `"claude-sonnet-4-5-20250929"`
963951
964 - `"claude-sonnet-4-5"`
965
966 High-performance model for agents and coding
967
968 - `"claude-sonnet-4-5-20250929"`
969
970 High-performance model for agents and coding
952 High-performance model for agents and coding
971953
972954 - `string`
973955
api/beta/agents/versions Changed · +158 / -166 lines
from line 37
3737
3838 - `string`
3939
40 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
40 - `"message-batches-2024-09-24"`
4141
42 - `"message-batches-2024-09-24"`
42 - `"prompt-caching-2024-07-31"`
4343
44 - `"prompt-caching-2024-07-31"`
44 - `"computer-use-2024-10-22"`
4545
46 - `"computer-use-2024-10-22"`
46 - `"computer-use-2025-01-24"`
4747
48 - `"computer-use-2025-01-24"`
48 - `"pdfs-2024-09-25"`
4949
50 - `"pdfs-2024-09-25"`
50 - `"token-counting-2024-11-01"`
5151
52 - `"token-counting-2024-11-01"`
52 - `"token-efficient-tools-2025-02-19"`
5353
54 - `"token-efficient-tools-2025-02-19"`
54 - `"output-128k-2025-02-19"`
5555
56 - `"output-128k-2025-02-19"`
56 - `"files-api-2025-04-14"`
5757
58 - `"files-api-2025-04-14"`
58 - `"mcp-client-2025-04-04"`
5959
60 - `"mcp-client-2025-04-04"`
60 - `"mcp-client-2025-11-20"`
6161
62 - `"mcp-client-2025-11-20"`
62 - `"dev-full-thinking-2025-05-14"`
6363
64 - `"dev-full-thinking-2025-05-14"`
64 - `"interleaved-thinking-2025-05-14"`
6565
66 - `"interleaved-thinking-2025-05-14"`
66 - `"code-execution-2025-05-22"`
6767
68 - `"code-execution-2025-05-22"`
68 - `"extended-cache-ttl-2025-04-11"`
6969
70 - `"extended-cache-ttl-2025-04-11"`
70 - `"context-1m-2025-08-07"`
7171
72 - `"context-1m-2025-08-07"`
72 - `"context-management-2025-06-27"`
7373
74 - `"context-management-2025-06-27"`
74 - `"model-context-window-exceeded-2025-08-26"`
7575
76 - `"model-context-window-exceeded-2025-08-26"`
76 - `"skills-2025-10-02"`
7777
78 - `"skills-2025-10-02"`
78 - `"fast-mode-2026-02-01"`
7979
80 - `"fast-mode-2026-02-01"`
80 - `"output-300k-2026-03-24"`
8181
82 - `"output-300k-2026-03-24"`
82 - `"user-profiles-2026-03-24"`
8383
84 - `"user-profiles-2026-03-24"`
84 - `"user-profiles-2026-08-18"`
8585
86 - `"user-profiles-2026-08-18"`
86 - `"user-profiles-2026-09-04"`
8787
88 - `"user-profiles-2026-09-04"`
88 - `"advisor-tool-2026-03-01"`
8989
90 - `"advisor-tool-2026-03-01"`
90 - `"managed-agents-2026-04-01"`
9191
92 - `"managed-agents-2026-04-01"`
92 - `"cache-diagnosis-2026-04-07"`
9393
94 - `"cache-diagnosis-2026-04-07"`
94 - `"dreaming-2026-04-21"`
9595
96 - `"dreaming-2026-04-21"`
96 - `"thinking-token-count-2026-05-13"`
9797
98 - `"thinking-token-count-2026-05-13"`
98 - `"server-side-fallback-2026-06-01"`
9999
100 - `"server-side-fallback-2026-06-01"`
100 - `"server-side-fallback-2026-07-01"`
101101
102 - `"server-side-fallback-2026-07-01"`
102 - `"fallback-credit-2026-06-01"`
103103
104 - `"fallback-credit-2026-06-01"`
104 - `"fallback-credit-2026-07-01"`
105105
106 - `"fallback-credit-2026-07-01"`
106 - `"agent-memory-2026-07-22"`
107107
108 - `"agent-memory-2026-07-22"`
108 - `"mid-conversation-tool-changes-2026-07-01"`
109109
110 - `"mid-conversation-tool-changes-2026-07-01"`
110 - `"compact-2026-01-12"`
111111
112 - `"compact-2026-01-12"`
112 - `"computer-use-2025-11-24"`
113113
114 - `"computer-use-2025-11-24"`
114 - `"mcp-tunnels-2026-06-22"`
115115
116 - `"mcp-tunnels-2026-06-22"`
116 - `"structured-outputs-2025-11-13"`
117117
118 - `"structured-outputs-2025-11-13"`
118 - `"task-budgets-2026-03-13"`
119119
120 - `"task-budgets-2026-03-13"`
120 - `"thinking-display-updates-2026-08-18"`
121121
122 - `"thinking-display-updates-2026-08-18"`
122 - `"ce-user-management-2026-07-13"`
123123
124 - `"ce-user-management-2026-07-13"`
124 - `"mid-conversation-output-config-2026-07-01"`
125125
126 - `"mid-conversation-output-config-2026-07-01"`
126 - `"thinking-binding-controls-2026-08-01"`
127127
128 - `"thinking-binding-controls-2026-08-01"`
128 - `"mid-conversation-system-clear-at-2026-08-21"`
129129
130 - `"mid-conversation-system-clear-at-2026-08-21"`
130 - `"compact-2026-09-04"`
131131
132 - `"compact-2026-09-04"`
132 - `"inline-tools-2026-09-15"`
133133
134 - `"inline-tools-2026-09-15"`
134 - `"mcp-client-2026-09-15"`
135135
136 - `"mcp-client-2026-09-15"`
137
138136- `"anthropic-workspace-id": optional string`
139137
140138 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 183
185183
186184 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
187185
188 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
186 - `"claude-sonnet-5-5"`
189187
190 The model that will power your agent.
188 Efficient model for coding and agents
191189
192 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
190 - `"claude-opus-5-5"`
193191
194 - `"claude-sonnet-5-5"`
192 Powerful intelligence for coding, knowledge work, and long-running agents
195193
196 Efficient model for coding and agents
194 - `"claude-fable-5-1"`
197195
198 - `"claude-opus-5-5"`
196 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
199197
200 Powerful intelligence for coding, knowledge work, and long-running agents
198 - `"claude-sonnet-5"`
201199
202 - `"claude-fable-5-1"`
200 Efficient model for coding and agents
203201
204 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
202 - `"claude-fable-5"`
205203
206 - `"claude-sonnet-5"`
204 Next generation of intelligence for the hardest knowledge work and coding problems
207205
208 Efficient model for coding and agents
206 - `"claude-opus-5"`
209207
210 - `"claude-fable-5"`
208 Powerful intelligence for long-running agents and coding
211209
212 Next generation of intelligence for the hardest knowledge work and coding problems
210 - `"claude-opus-4-8"`
213211
214 - `"claude-opus-5"`
212 Powerful intelligence for long-running agents and coding
215213
216 Powerful intelligence for long-running agents and coding
214 - `"claude-opus-4-7"`
217215
218 - `"claude-opus-4-8"`
216 Powerful intelligence for long-running agents and coding
219217
220 Powerful intelligence for long-running agents and coding
218 - `"claude-opus-4-6"`
221219
222 - `"claude-opus-4-7"`
220 Powerful intelligence for long-running agents and coding
223221
224 Powerful intelligence for long-running agents and coding
222 - `"claude-sonnet-4-6"`
225223
226 - `"claude-opus-4-6"`
224 Best combination of speed and intelligence
227225
228 Powerful intelligence for long-running agents and coding
226 - `"claude-haiku-4-5"`
229227
230 - `"claude-sonnet-4-6"`
228 Fastest model with near-frontier intelligence
231229
232 Best combination of speed and intelligence
230 - `"claude-haiku-4-5-20251001"`
233231
234 - `"claude-haiku-4-5"`
232 Fastest model with near-frontier intelligence
235233
236 Fastest model with near-frontier intelligence
234 - `"claude-opus-4-5"`
237235
238 - `"claude-haiku-4-5-20251001"`
236 Powerful intelligence for long-running agents and coding
239237
240 Fastest model with near-frontier intelligence
238 - `"claude-opus-4-5-20251101"`
241239
242 - `"claude-opus-4-5"`
240 Powerful intelligence for long-running agents and coding
243241
244 Powerful intelligence for long-running agents and coding
242 - `"claude-sonnet-4-5"`
245243
246 - `"claude-opus-4-5-20251101"`
244 High-performance model for agents and coding
247245
248 Powerful intelligence for long-running agents and coding
246 - `"claude-sonnet-4-5-20250929"`
249247
250 - `"claude-sonnet-4-5"`
251
252 High-performance model for agents and coding
253
254 - `"claude-sonnet-4-5-20250929"`
255
256 High-performance model for agents and coding
248 High-performance model for agents and coding
257249
258250 - `string`
259251
api/beta/agents/versions/list Changed · +158 / -166 lines
from line 35
3535
3636 - `string`
3737
38 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
38 - `"message-batches-2024-09-24"`
3939
40 - `"message-batches-2024-09-24"`
40 - `"prompt-caching-2024-07-31"`
4141
42 - `"prompt-caching-2024-07-31"`
42 - `"computer-use-2024-10-22"`
4343
44 - `"computer-use-2024-10-22"`
44 - `"computer-use-2025-01-24"`
4545
46 - `"computer-use-2025-01-24"`
46 - `"pdfs-2024-09-25"`
4747
48 - `"pdfs-2024-09-25"`
48 - `"token-counting-2024-11-01"`
4949
50 - `"token-counting-2024-11-01"`
50 - `"token-efficient-tools-2025-02-19"`
5151
52 - `"token-efficient-tools-2025-02-19"`
52 - `"output-128k-2025-02-19"`
5353
54 - `"output-128k-2025-02-19"`
54 - `"files-api-2025-04-14"`
5555
56 - `"files-api-2025-04-14"`
56 - `"mcp-client-2025-04-04"`
5757
58 - `"mcp-client-2025-04-04"`
58 - `"mcp-client-2025-11-20"`
5959
60 - `"mcp-client-2025-11-20"`
60 - `"dev-full-thinking-2025-05-14"`
6161
62 - `"dev-full-thinking-2025-05-14"`
62 - `"interleaved-thinking-2025-05-14"`
6363
64 - `"interleaved-thinking-2025-05-14"`
64 - `"code-execution-2025-05-22"`
6565
66 - `"code-execution-2025-05-22"`
66 - `"extended-cache-ttl-2025-04-11"`
6767
68 - `"extended-cache-ttl-2025-04-11"`
68 - `"context-1m-2025-08-07"`
6969
70 - `"context-1m-2025-08-07"`
70 - `"context-management-2025-06-27"`
7171
72 - `"context-management-2025-06-27"`
72 - `"model-context-window-exceeded-2025-08-26"`
7373
74 - `"model-context-window-exceeded-2025-08-26"`
74 - `"skills-2025-10-02"`
7575
76 - `"skills-2025-10-02"`
76 - `"fast-mode-2026-02-01"`
7777
78 - `"fast-mode-2026-02-01"`
78 - `"output-300k-2026-03-24"`
7979
80 - `"output-300k-2026-03-24"`
80 - `"user-profiles-2026-03-24"`
8181
82 - `"user-profiles-2026-03-24"`
82 - `"user-profiles-2026-08-18"`
8383
84 - `"user-profiles-2026-08-18"`
84 - `"user-profiles-2026-09-04"`
8585
86 - `"user-profiles-2026-09-04"`
86 - `"advisor-tool-2026-03-01"`
8787
88 - `"advisor-tool-2026-03-01"`
88 - `"managed-agents-2026-04-01"`
8989
90 - `"managed-agents-2026-04-01"`
90 - `"cache-diagnosis-2026-04-07"`
9191
92 - `"cache-diagnosis-2026-04-07"`
92 - `"dreaming-2026-04-21"`
9393
94 - `"dreaming-2026-04-21"`
94 - `"thinking-token-count-2026-05-13"`
9595
96 - `"thinking-token-count-2026-05-13"`
96 - `"server-side-fallback-2026-06-01"`
9797
98 - `"server-side-fallback-2026-06-01"`
98 - `"server-side-fallback-2026-07-01"`
9999
100 - `"server-side-fallback-2026-07-01"`
100 - `"fallback-credit-2026-06-01"`
101101
102 - `"fallback-credit-2026-06-01"`
102 - `"fallback-credit-2026-07-01"`
103103
104 - `"fallback-credit-2026-07-01"`
104 - `"agent-memory-2026-07-22"`
105105
106 - `"agent-memory-2026-07-22"`
106 - `"mid-conversation-tool-changes-2026-07-01"`
107107
108 - `"mid-conversation-tool-changes-2026-07-01"`
108 - `"compact-2026-01-12"`
109109
110 - `"compact-2026-01-12"`
110 - `"computer-use-2025-11-24"`
111111
112 - `"computer-use-2025-11-24"`
112 - `"mcp-tunnels-2026-06-22"`
113113
114 - `"mcp-tunnels-2026-06-22"`
114 - `"structured-outputs-2025-11-13"`
115115
116 - `"structured-outputs-2025-11-13"`
116 - `"task-budgets-2026-03-13"`
117117
118 - `"task-budgets-2026-03-13"`
118 - `"thinking-display-updates-2026-08-18"`
119119
120 - `"thinking-display-updates-2026-08-18"`
120 - `"ce-user-management-2026-07-13"`
121121
122 - `"ce-user-management-2026-07-13"`
122 - `"mid-conversation-output-config-2026-07-01"`
123123
124 - `"mid-conversation-output-config-2026-07-01"`
124 - `"thinking-binding-controls-2026-08-01"`
125125
126 - `"thinking-binding-controls-2026-08-01"`
126 - `"mid-conversation-system-clear-at-2026-08-21"`
127127
128 - `"mid-conversation-system-clear-at-2026-08-21"`
128 - `"compact-2026-09-04"`
129129
130 - `"compact-2026-09-04"`
130 - `"inline-tools-2026-09-15"`
131131
132 - `"inline-tools-2026-09-15"`
132 - `"mcp-client-2026-09-15"`
133133
134 - `"mcp-client-2026-09-15"`
135
136134- `"anthropic-workspace-id": optional string`
137135
138136 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 181
183181
184182 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
185183
186 - `"claude-sonnet-5-5" or "claude-opus-5-5" or "claude-fable-5-1" or 13 more`
184 - `"claude-sonnet-5-5"`
187185
188 The model that will power your agent.
186 Efficient model for coding and agents
189187
190 See [models](https://docs.anthropic.com/en/docs/models-overview) for additional details and options.
188 - `"claude-opus-5-5"`
191189
192 - `"claude-sonnet-5-5"`
190 Powerful intelligence for coding, knowledge work, and long-running agents
193191
194 Efficient model for coding and agents
192 - `"claude-fable-5-1"`
195193
196 - `"claude-opus-5-5"`
194 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
197195
198 Powerful intelligence for coding, knowledge work, and long-running agents
196 - `"claude-sonnet-5"`
199197
200 - `"claude-fable-5-1"`
198 Efficient model for coding and agents
201199
202 Frontier intelligence for ambitious tasks across coding, scientific discovery, and enterprise workflows
200 - `"claude-fable-5"`
203201
204 - `"claude-sonnet-5"`
202 Next generation of intelligence for the hardest knowledge work and coding problems
205203
206 Efficient model for coding and agents
204 - `"claude-opus-5"`
207205
208 - `"claude-fable-5"`
206 Powerful intelligence for long-running agents and coding
209207
210 Next generation of intelligence for the hardest knowledge work and coding problems
208 - `"claude-opus-4-8"`
211209
212 - `"claude-opus-5"`
210 Powerful intelligence for long-running agents and coding
213211
214 Powerful intelligence for long-running agents and coding
212 - `"claude-opus-4-7"`
215213
216 - `"claude-opus-4-8"`
214 Powerful intelligence for long-running agents and coding
217215
218 Powerful intelligence for long-running agents and coding
216 - `"claude-opus-4-6"`
219217
220 - `"claude-opus-4-7"`
218 Powerful intelligence for long-running agents and coding
221219
222 Powerful intelligence for long-running agents and coding
220 - `"claude-sonnet-4-6"`
223221
224 - `"claude-opus-4-6"`
222 Best combination of speed and intelligence
225223
226 Powerful intelligence for long-running agents and coding
224 - `"claude-haiku-4-5"`
227225
228 - `"claude-sonnet-4-6"`
226 Fastest model with near-frontier intelligence
229227
230 Best combination of speed and intelligence
228 - `"claude-haiku-4-5-20251001"`
231229
232 - `"claude-haiku-4-5"`
230 Fastest model with near-frontier intelligence
233231
234 Fastest model with near-frontier intelligence
232 - `"claude-opus-4-5"`
235233
236 - `"claude-haiku-4-5-20251001"`
234 Powerful intelligence for long-running agents and coding
237235
238 Fastest model with near-frontier intelligence
236 - `"claude-opus-4-5-20251101"`
239237
240 - `"claude-opus-4-5"`
238 Powerful intelligence for long-running agents and coding
241239
242 Powerful intelligence for long-running agents and coding
240 - `"claude-sonnet-4-5"`
243241
244 - `"claude-opus-4-5-20251101"`
242 High-performance model for agents and coding
245243
246 Powerful intelligence for long-running agents and coding
244 - `"claude-sonnet-4-5-20250929"`
247245
248 - `"claude-sonnet-4-5"`
249
250 High-performance model for agents and coding
251
252 - `"claude-sonnet-4-5-20250929"`
253
254 High-performance model for agents and coding
246 High-performance model for agents and coding
255247
256248 - `string`
257249
api/beta/deployment_runs Changed · +190 / -194 lines
from line 75
7575
7676 - `string`
7777
78 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
78 - `"message-batches-2024-09-24"`
7979
80 - `"message-batches-2024-09-24"`
80 - `"prompt-caching-2024-07-31"`
8181
82 - `"prompt-caching-2024-07-31"`
82 - `"computer-use-2024-10-22"`
8383
84 - `"computer-use-2024-10-22"`
84 - `"computer-use-2025-01-24"`
8585
86 - `"computer-use-2025-01-24"`
86 - `"pdfs-2024-09-25"`
8787
88 - `"pdfs-2024-09-25"`
88 - `"token-counting-2024-11-01"`
8989
90 - `"token-counting-2024-11-01"`
90 - `"token-efficient-tools-2025-02-19"`
9191
92 - `"token-efficient-tools-2025-02-19"`
92 - `"output-128k-2025-02-19"`
9393
94 - `"output-128k-2025-02-19"`
94 - `"files-api-2025-04-14"`
9595
96 - `"files-api-2025-04-14"`
96 - `"mcp-client-2025-04-04"`
9797
98 - `"mcp-client-2025-04-04"`
98 - `"mcp-client-2025-11-20"`
9999
100 - `"mcp-client-2025-11-20"`
100 - `"dev-full-thinking-2025-05-14"`
101101
102 - `"dev-full-thinking-2025-05-14"`
102 - `"interleaved-thinking-2025-05-14"`
103103
104 - `"interleaved-thinking-2025-05-14"`
104 - `"code-execution-2025-05-22"`
105105
106 - `"code-execution-2025-05-22"`
106 - `"extended-cache-ttl-2025-04-11"`
107107
108 - `"extended-cache-ttl-2025-04-11"`
108 - `"context-1m-2025-08-07"`
109109
110 - `"context-1m-2025-08-07"`
110 - `"context-management-2025-06-27"`
111111
112 - `"context-management-2025-06-27"`
112 - `"model-context-window-exceeded-2025-08-26"`
113113
114 - `"model-context-window-exceeded-2025-08-26"`
114 - `"skills-2025-10-02"`
115115
116 - `"skills-2025-10-02"`
116 - `"fast-mode-2026-02-01"`
117117
118 - `"fast-mode-2026-02-01"`
118 - `"output-300k-2026-03-24"`
119119
120 - `"output-300k-2026-03-24"`
120 - `"user-profiles-2026-03-24"`
121121
122 - `"user-profiles-2026-03-24"`
122 - `"user-profiles-2026-08-18"`
123123
124 - `"user-profiles-2026-08-18"`
124 - `"user-profiles-2026-09-04"`
125125
126 - `"user-profiles-2026-09-04"`
126 - `"advisor-tool-2026-03-01"`
127127
128 - `"advisor-tool-2026-03-01"`
128 - `"managed-agents-2026-04-01"`
129129
130 - `"managed-agents-2026-04-01"`
130 - `"cache-diagnosis-2026-04-07"`
131131
132 - `"cache-diagnosis-2026-04-07"`
132 - `"dreaming-2026-04-21"`
133133
134 - `"dreaming-2026-04-21"`
134 - `"thinking-token-count-2026-05-13"`
135135
136 - `"thinking-token-count-2026-05-13"`
136 - `"server-side-fallback-2026-06-01"`
137137
138 - `"server-side-fallback-2026-06-01"`
138 - `"server-side-fallback-2026-07-01"`
139139
140 - `"server-side-fallback-2026-07-01"`
140 - `"fallback-credit-2026-06-01"`
141141
142 - `"fallback-credit-2026-06-01"`
142 - `"fallback-credit-2026-07-01"`
143143
144 - `"fallback-credit-2026-07-01"`
144 - `"agent-memory-2026-07-22"`
145145
146 - `"agent-memory-2026-07-22"`
146 - `"mid-conversation-tool-changes-2026-07-01"`
147147
148 - `"mid-conversation-tool-changes-2026-07-01"`
148 - `"compact-2026-01-12"`
149149
150 - `"compact-2026-01-12"`
150 - `"computer-use-2025-11-24"`
151151
152 - `"computer-use-2025-11-24"`
152 - `"mcp-tunnels-2026-06-22"`
153153
154 - `"mcp-tunnels-2026-06-22"`
154 - `"structured-outputs-2025-11-13"`
155155
156 - `"structured-outputs-2025-11-13"`
156 - `"task-budgets-2026-03-13"`
157157
158 - `"task-budgets-2026-03-13"`
158 - `"thinking-display-updates-2026-08-18"`
159159
160 - `"thinking-display-updates-2026-08-18"`
160 - `"ce-user-management-2026-07-13"`
161161
162 - `"ce-user-management-2026-07-13"`
162 - `"mid-conversation-output-config-2026-07-01"`
163163
164 - `"mid-conversation-output-config-2026-07-01"`
164 - `"thinking-binding-controls-2026-08-01"`
165165
166 - `"thinking-binding-controls-2026-08-01"`
166 - `"mid-conversation-system-clear-at-2026-08-21"`
167167
168 - `"mid-conversation-system-clear-at-2026-08-21"`
168 - `"compact-2026-09-04"`
169169
170 - `"compact-2026-09-04"`
170 - `"inline-tools-2026-09-15"`
171171
172 - `"inline-tools-2026-09-15"`
172 - `"mcp-client-2026-09-15"`
173173
174 - `"mcp-client-2026-09-15"`
175
176174- `"anthropic-workspace-id": optional string`
177175
178176 Optional header to select the Workspace for this request. The value is a Workspace ID (for example, `wrkspc_011CZkZaBF1tNoB5wlCeusgy`).
from line 464
466464
467465 - `string`
468466
469 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
467 - `"message-batches-2024-09-24"`
470468
471 - `"message-batches-2024-09-24"`
469 - `"prompt-caching-2024-07-31"`
472470
473 - `"prompt-caching-2024-07-31"`
471 - `"computer-use-2024-10-22"`
474472
475 - `"computer-use-2024-10-22"`
473 - `"computer-use-2025-01-24"`
476474
477 - `"computer-use-2025-01-24"`
475 - `"pdfs-2024-09-25"`
478476
479 - `"pdfs-2024-09-25"`
477 - `"token-counting-2024-11-01"`
480478
481 - `"token-counting-2024-11-01"`
479 - `"token-efficient-tools-2025-02-19"`
482480
483 - `"token-efficient-tools-2025-02-19"`
481 - `"output-128k-2025-02-19"`
484482
485 - `"output-128k-2025-02-19"`
483 - `"files-api-2025-04-14"`
486484
487 - `"files-api-2025-04-14"`
485 - `"mcp-client-2025-04-04"`
488486
489 - `"mcp-client-2025-04-04"`
487 - `"mcp-client-2025-11-20"`
490488
491 - `"mcp-client-2025-11-20"`
489 - `"dev-full-thinking-2025-05-14"`
492490
493 - `"dev-full-thinking-2025-05-14"`
491 - `"interleaved-thinking-2025-05-14"`
494492
495 - `"interleaved-thinking-2025-05-14"`
493 - `"code-execution-2025-05-22"`
496494
497 - `"code-execution-2025-05-22"`
495 - `"extended-cache-ttl-2025-04-11"`
498496
499 - `"extended-cache-ttl-2025-04-11"`
497 - `"context-1m-2025-08-07"`
500498
501 - `"context-1m-2025-08-07"`
499 - `"context-management-2025-06-27"`
502500
503 - `"context-management-2025-06-27"`
501 - `"model-context-window-exceeded-2025-08-26"`
504502
505 - `"model-context-window-exceeded-2025-08-26"`
503 - `"skills-2025-10-02"`
506504
507 - `"skills-2025-10-02"`
505 - `"fast-mode-2026-02-01"`
508506
509 - `"fast-mode-2026-02-01"`
507 - `"output-300k-2026-03-24"`
510508
511 - `"output-300k-2026-03-24"`
509 - `"user-profiles-2026-03-24"`
512510
513 - `"user-profiles-2026-03-24"`
511 - `"user-profiles-2026-08-18"`
514512
515 - `"user-profiles-2026-08-18"`
513 - `"user-profiles-2026-09-04"`
516514
517 - `"user-profiles-2026-09-04"`
515 - `"advisor-tool-2026-03-01"`
518516
519 - `"advisor-tool-2026-03-01"`
517 - `"managed-agents-2026-04-01"`
520518
521 - `"managed-agents-2026-04-01"`
519 - `"cache-diagnosis-2026-04-07"`
522520
523 - `"cache-diagnosis-2026-04-07"`
521 - `"dreaming-2026-04-21"`
524522
525 - `"dreaming-2026-04-21"`
523 - `"thinking-token-count-2026-05-13"`
526524
527 - `"thinking-token-count-2026-05-13"`
525 - `"server-side-fallback-2026-06-01"`
528526
529 - `"server-side-fallback-2026-06-01"`
527 - `"server-side-fallback-2026-07-01"`
530528
531 - `"server-side-fallback-2026-07-01"`
529 - `"fallback-credit-2026-06-01"`
532530
533 - `"fallback-credit-2026-06-01"`
531 - `"fallback-credit-2026-07-01"`
534532
535 - `"fallback-credit-2026-07-01"`
533 - `"agent-memory-2026-07-22"`
536534
537 - `"agent-memory-2026-07-22"`
535 - `"mid-conversation-tool-changes-2026-07-01"`
538536
539 - `"mid-conversation-tool-changes-2026-07-01"`
537 - `"compact-2026-01-12"`
540538
541 - `"compact-2026-01-12"`
539 - `"computer-use-2025-11-24"`
542540
543 - `"computer-use-2025-11-24"`
541 - `"mcp-tunnels-2026-06-22"`
544542
545 - `"mcp-tunnels-2026-06-22"`
543 - `"structured-outputs-2025-11-13"`
546544
547 - `"structured-outputs-2025-11-13"`
545 - `"task-budgets-2026-03-13"`
548546
549 - `"task-budgets-2026-03-13"`
547 - `"thinking-display-updates-2026-08-18"`
550548
551 - `"thinking-display-updates-2026-08-18"`
549 - `"ce-user-management-2026-07-13"`
552550
553 - `"ce-user-management-2026-07-13"`
551 - `"mid-conversation-output-config-2026-07-01"`
554552
555 - `"mid-conversation-output-config-2026-07-01"`
553 - `"thinking-binding-controls-2026-08-01"`
556554
557 - `"thinking-binding-controls-2026-08-01"`
555 - `"mid-conversation-system-clear-at-2026-08-21"`
558556
559 - `"mid-conversation-system-clear-at-2026-08-21"`
557 - `"compact-2026-09-04"`
560558
561 - `"compact-2026-09-04"`
559 - `"inline-tools-2026-09-15"`
562560
563 - `"inline-tools-2026-09-15"`
564
565 - `"mcp-client-2026-09-15"`
561 - `"mcp-client-2026-09-15"`
566562
567563- `"anthropic-workspace-id": optional string`
568564
api/beta/deployment_runs/list Changed · +95 / -97 lines
from line 73
7373
7474 - `string`
7575
76 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
76 - `"message-batches-2024-09-24"`
7777
78 - `"message-batches-2024-09-24"`
78 - `"prompt-caching-2024-07-31"`
7979
80 - `"prompt-caching-2024-07-31"`
80 - `"computer-use-2024-10-22"`
8181
82 - `"computer-use-2024-10-22"`
82 - `"computer-use-2025-01-24"`
8383
84 - `"computer-use-2025-01-24"`
84 - `"pdfs-2024-09-25"`
8585
86 - `"pdfs-2024-09-25"`
86 - `"token-counting-2024-11-01"`
8787
88 - `"token-counting-2024-11-01"`
88 - `"token-efficient-tools-2025-02-19"`
8989
90 - `"token-efficient-tools-2025-02-19"`
90 - `"output-128k-2025-02-19"`
9191
92 - `"output-128k-2025-02-19"`
92 - `"files-api-2025-04-14"`
9393
94 - `"files-api-2025-04-14"`
94 - `"mcp-client-2025-04-04"`
9595
96 - `"mcp-client-2025-04-04"`
96 - `"mcp-client-2025-11-20"`
9797
98 - `"mcp-client-2025-11-20"`
98 - `"dev-full-thinking-2025-05-14"`
9999
100 - `"dev-full-thinking-2025-05-14"`
100 - `"interleaved-thinking-2025-05-14"`
101101
102 - `"interleaved-thinking-2025-05-14"`
102 - `"code-execution-2025-05-22"`
103103
104 - `"code-execution-2025-05-22"`
104 - `"extended-cache-ttl-2025-04-11"`
105105
106 - `"extended-cache-ttl-2025-04-11"`
106 - `"context-1m-2025-08-07"`
107107
108 - `"context-1m-2025-08-07"`
108 - `"context-management-2025-06-27"`
109109
110 - `"context-management-2025-06-27"`
110 - `"model-context-window-exceeded-2025-08-26"`
111111
112 - `"model-context-window-exceeded-2025-08-26"`
112 - `"skills-2025-10-02"`
113113
114 - `"skills-2025-10-02"`
114 - `"fast-mode-2026-02-01"`
115115
116 - `"fast-mode-2026-02-01"`
116 - `"output-300k-2026-03-24"`
117117
118 - `"output-300k-2026-03-24"`
118 - `"user-profiles-2026-03-24"`
119119
120 - `"user-profiles-2026-03-24"`
120 - `"user-profiles-2026-08-18"`
121121
122 - `"user-profiles-2026-08-18"`
122 - `"user-profiles-2026-09-04"`
123123
124 - `"user-profiles-2026-09-04"`
124 - `"advisor-tool-2026-03-01"`
125125
126 - `"advisor-tool-2026-03-01"`
126 - `"managed-agents-2026-04-01"`
127127
128 - `"managed-agents-2026-04-01"`
128 - `"cache-diagnosis-2026-04-07"`
129129
130 - `"cache-diagnosis-2026-04-07"`
130 - `"dreaming-2026-04-21"`
131131
132 - `"dreaming-2026-04-21"`
132 - `"thinking-token-count-2026-05-13"`
133133
134 - `"thinking-token-count-2026-05-13"`
134 - `"server-side-fallback-2026-06-01"`
135135
136 - `"server-side-fallback-2026-06-01"`
136 - `"server-side-fallback-2026-07-01"`
137137
138 - `"server-side-fallback-2026-07-01"`
138 - `"fallback-credit-2026-06-01"`
139139
140 - `"fallback-credit-2026-06-01"`
140 - `"fallback-credit-2026-07-01"`
141141
142 - `"fallback-credit-2026-07-01"`
142 - `"agent-memory-2026-07-22"`
143143
144 - `"agent-memory-2026-07-22"`
144 - `"mid-conversation-tool-changes-2026-07-01"`
145145
146 - `"mid-conversation-tool-changes-2026-07-01"`
146 - `"compact-2026-01-12"`
147147
148 - `"compact-2026-01-12"`
148 - `"computer-use-2025-11-24"`
149149
150 - `"computer-use-2025-11-24"`
150 - `"mcp-tunnels-2026-06-22"`
151151
152 - `"mcp-tunnels-2026-06-22"`
152 - `"structured-outputs-2025-11-13"`
153153
154 - `"structured-outputs-2025-11-13"`
154 - `"task-budgets-2026-03-13"`
155155
156 - `"task-budgets-2026-03-13"`
156 - `"thinking-display-updates-2026-08-18"`
157157
158 - `"thinking-display-updates-2026-08-18"`
158 - `"ce-user-management-2026-07-13"`
159159
160 - `"ce-user-management-2026-07-13"`
160 - `"mid-conversation-output-config-2026-07-01"`
161161
162 - `"mid-conversation-output-config-2026-07-01"`
162 - `"thinking-binding-controls-2026-08-01"`
163163
164 - `"thinking-binding-controls-2026-08-01"`
164 - `"mid-conversation-system-clear-at-2026-08-21"`
165165
166 - `"mid-conversation-system-clear-at-2026-08-21"`
166 - `"compact-2026-09-04"`
167167
168 - `"compact-2026-09-04"`
168 - `"inline-tools-2026-09-15"`
169169
170 - `"inline-tools-2026-09-15"`
171
172 - `"mcp-client-2026-09-15"`
170 - `"mcp-client-2026-09-15"`
173171
174172- `"anthropic-workspace-id": optional string`
175173
api/beta/deployment_runs/retrieve Changed · +95 / -97 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
121
122 - `"mcp-client-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
123121
124122- `"anthropic-workspace-id": optional string`
125123
api/beta/deployments Changed · +760 / -776 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 19
1919
2020 - `string`
2121
22 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
23
24 - `"message-batches-2024-09-24"`
25
26 - `"prompt-caching-2024-07-31"`
27
28 - `"computer-use-2024-10-22"`
29
30 - `"computer-use-2025-01-24"`
31
32 - `"pdfs-2024-09-25"`
33
34 - `"token-counting-2024-11-01"`
35
36 - `"token-efficient-tools-2025-02-19"`
37
38 - `"output-128k-2025-02-19"`
39
40 - `"files-api-2025-04-14"`
41
42 - `"mcp-client-2025-04-04"`
43
44 - `"mcp-client-2025-11-20"`
45
46 - `"dev-full-thinking-2025-05-14"`
47
48 - `"interleaved-thinking-2025-05-14"`
49
50 - `"code-execution-2025-05-22"`
51
52 - `"extended-cache-ttl-2025-04-11"`
53
54 - `"context-1m-2025-08-07"`
55
56 - `"context-management-2025-06-27"`
57
58 - `"model-context-window-exceeded-2025-08-26"`
59
60 - `"skills-2025-10-02"`
61
62 - `"fast-mode-2026-02-01"`
63
64 - `"output-300k-2026-03-24"`
65
66 - `"user-profiles-2026-03-24"`
67
68 - `"user-profiles-2026-08-18"`
69
70 - `"user-profiles-2026-09-04"`
71
72 - `"advisor-tool-2026-03-01"`
73
74 - `"managed-agents-2026-04-01"`
75
76 - `"cache-diagnosis-2026-04-07"`
77
78 - `"dreaming-2026-04-21"`
79
80 - `"thinking-token-count-2026-05-13"`
81
82 - `"server-side-fallback-2026-06-01"`
83
84 - `"server-side-fallback-2026-07-01"`
85
86 - `"fallback-credit-2026-06-01"`
87
88 - `"fallback-credit-2026-07-01"`
89
90 - `"agent-memory-2026-07-22"`
91
92 - `"mid-conversation-tool-changes-2026-07-01"`
93
94 - `"compact-2026-01-12"`
95
96 - `"computer-use-2025-11-24"`
97
98 - `"mcp-tunnels-2026-06-22"`
99
100 - `"structured-outputs-2025-11-13"`
101
102 - `"task-budgets-2026-03-13"`
103
104 - `"thinking-display-updates-2026-08-18"`
105
106 - `"ce-user-management-2026-07-13"`
107
108 - `"mid-conversation-output-config-2026-07-01"`
109
110 - `"thinking-binding-controls-2026-08-01"`
111
112 - `"mid-conversation-system-clear-at-2026-08-21"`
113
114 - `"compact-2026-09-04"`
115
116 - `"inline-tools-2026-09-15"`
117
118 - `"mcp-client-2026-09-15"`
22 - `"message-batches-2024-09-24"`
23
24 - `"prompt-caching-2024-07-31"`
25
26 - `"computer-use-2024-10-22"`
27
28 - `"computer-use-2025-01-24"`
29
30 - `"pdfs-2024-09-25"`
31
32 - `"token-counting-2024-11-01"`
33
34 - `"token-efficient-tools-2025-02-19"`
35
36 - `"output-128k-2025-02-19"`
37
38 - `"files-api-2025-04-14"`
39
40 - `"mcp-client-2025-04-04"`
41
42 - `"mcp-client-2025-11-20"`
43
44 - `"dev-full-thinking-2025-05-14"`
45
46 - `"interleaved-thinking-2025-05-14"`
47
48 - `"code-execution-2025-05-22"`
49
50 - `"extended-cache-ttl-2025-04-11"`
51
52 - `"context-1m-2025-08-07"`
53
54 - `"context-management-2025-06-27"`
55
56 - `"model-context-window-exceeded-2025-08-26"`
57
58 - `"skills-2025-10-02"`
59
60 - `"fast-mode-2026-02-01"`
61
62 - `"output-300k-2026-03-24"`
63
64 - `"user-profiles-2026-03-24"`
65
66 - `"user-profiles-2026-08-18"`
67
68 - `"user-profiles-2026-09-04"`
69
70 - `"advisor-tool-2026-03-01"`
71
72 - `"managed-agents-2026-04-01"`
73
74 - `"cache-diagnosis-2026-04-07"`
75
76 - `"dreaming-2026-04-21"`
77
78 - `"thinking-token-count-2026-05-13"`
79
80 - `"server-side-fallback-2026-06-01"`
81
82 - `"server-side-fallback-2026-07-01"`
83
84 - `"fallback-credit-2026-06-01"`
85
86 - `"fallback-credit-2026-07-01"`
87
88 - `"agent-memory-2026-07-22"`
89
90 - `"mid-conversation-tool-changes-2026-07-01"`
91
92 - `"compact-2026-01-12"`
93
94 - `"computer-use-2025-11-24"`
95
96 - `"mcp-tunnels-2026-06-22"`
97
98 - `"structured-outputs-2025-11-13"`
99
100 - `"task-budgets-2026-03-13"`
101
102 - `"thinking-display-updates-2026-08-18"`
103
104 - `"ce-user-management-2026-07-13"`
105
106 - `"mid-conversation-output-config-2026-07-01"`
107
108 - `"thinking-binding-controls-2026-08-01"`
109
110 - `"mid-conversation-system-clear-at-2026-08-21"`
111
112 - `"compact-2026-09-04"`
113
114 - `"inline-tools-2026-09-15"`
115
116 - `"mcp-client-2026-09-15"`
119117
120118- `"anthropic-workspace-id": optional string`
121119
from line 1199
12011199
12021200 - `string`
12031201
1204 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
1205
1206 - `"message-batches-2024-09-24"`
1207
1208 - `"prompt-caching-2024-07-31"`
1209
1210 - `"computer-use-2024-10-22"`
1211
1212 - `"computer-use-2025-01-24"`
1213
1214 - `"pdfs-2024-09-25"`
1215
1216 - `"token-counting-2024-11-01"`
1217
1218 - `"token-efficient-tools-2025-02-19"`
1219
1220 - `"output-128k-2025-02-19"`
1221
1222 - `"files-api-2025-04-14"`
1223
1224 - `"mcp-client-2025-04-04"`
1225
1226 - `"mcp-client-2025-11-20"`
1227
1228 - `"dev-full-thinking-2025-05-14"`
1229
1230 - `"interleaved-thinking-2025-05-14"`
1231
1232 - `"code-execution-2025-05-22"`
1233
1234 - `"extended-cache-ttl-2025-04-11"`
1235
1236 - `"context-1m-2025-08-07"`
1237
1238 - `"context-management-2025-06-27"`
1239
1240 - `"model-context-window-exceeded-2025-08-26"`
1241
1242 - `"skills-2025-10-02"`
1243
1244 - `"fast-mode-2026-02-01"`
1245
1246 - `"output-300k-2026-03-24"`
1247
1248 - `"user-profiles-2026-03-24"`
1249
1250 - `"user-profiles-2026-08-18"`
1251
1252 - `"user-profiles-2026-09-04"`
1253
1254 - `"advisor-tool-2026-03-01"`
1255
1256 - `"managed-agents-2026-04-01"`
1257
1258 - `"cache-diagnosis-2026-04-07"`
1259
1260 - `"dreaming-2026-04-21"`
1261
1262 - `"thinking-token-count-2026-05-13"`
1263
1264 - `"server-side-fallback-2026-06-01"`
1265
1266 - `"server-side-fallback-2026-07-01"`
1267
1268 - `"fallback-credit-2026-06-01"`
1269
1270 - `"fallback-credit-2026-07-01"`
1271
1272 - `"agent-memory-2026-07-22"`
1273
1274 - `"mid-conversation-tool-changes-2026-07-01"`
1275
1276 - `"compact-2026-01-12"`
1277
1278 - `"computer-use-2025-11-24"`
1279
1280 - `"mcp-tunnels-2026-06-22"`
1281
1282 - `"structured-outputs-2025-11-13"`
1283
1284 - `"task-budgets-2026-03-13"`
1285
1286 - `"thinking-display-updates-2026-08-18"`
1287
1288 - `"ce-user-management-2026-07-13"`
1289
1290 - `"mid-conversation-output-config-2026-07-01"`
1291
1292 - `"thinking-binding-controls-2026-08-01"`
1293
1294 - `"mid-conversation-system-clear-at-2026-08-21"`
1295
1296 - `"compact-2026-09-04"`
1297
1298 - `"inline-tools-2026-09-15"`
1299
1300 - `"mcp-client-2026-09-15"`
1202 - `"message-batches-2024-09-24"`
1203
1204 - `"prompt-caching-2024-07-31"`
1205
1206 - `"computer-use-2024-10-22"`
1207
1208 - `"computer-use-2025-01-24"`
1209
1210 - `"pdfs-2024-09-25"`
1211
1212 - `"token-counting-2024-11-01"`
1213
1214 - `"token-efficient-tools-2025-02-19"`
1215
1216 - `"output-128k-2025-02-19"`
1217
1218 - `"files-api-2025-04-14"`
1219
1220 - `"mcp-client-2025-04-04"`
1221
1222 - `"mcp-client-2025-11-20"`
1223
1224 - `"dev-full-thinking-2025-05-14"`
1225
1226 - `"interleaved-thinking-2025-05-14"`
1227
1228 - `"code-execution-2025-05-22"`
1229
1230 - `"extended-cache-ttl-2025-04-11"`
1231
1232 - `"context-1m-2025-08-07"`
1233
1234 - `"context-management-2025-06-27"`
1235
1236 - `"model-context-window-exceeded-2025-08-26"`
1237
1238 - `"skills-2025-10-02"`
1239
1240 - `"fast-mode-2026-02-01"`
1241
1242 - `"output-300k-2026-03-24"`
1243
1244 - `"user-profiles-2026-03-24"`
1245
1246 - `"user-profiles-2026-08-18"`
1247
1248 - `"user-profiles-2026-09-04"`
1249
1250 - `"advisor-tool-2026-03-01"`
1251
1252 - `"managed-agents-2026-04-01"`
1253
1254 - `"cache-diagnosis-2026-04-07"`
1255
1256 - `"dreaming-2026-04-21"`
1257
1258 - `"thinking-token-count-2026-05-13"`
1259
1260 - `"server-side-fallback-2026-06-01"`
1261
1262 - `"server-side-fallback-2026-07-01"`
1263
1264 - `"fallback-credit-2026-06-01"`
1265
1266 - `"fallback-credit-2026-07-01"`
1267
1268 - `"agent-memory-2026-07-22"`
1269
1270 - `"mid-conversation-tool-changes-2026-07-01"`
1271
1272 - `"compact-2026-01-12"`
1273
1274 - `"computer-use-2025-11-24"`
1275
1276 - `"mcp-tunnels-2026-06-22"`
1277
1278 - `"structured-outputs-2025-11-13"`
1279
1280 - `"task-budgets-2026-03-13"`
1281
1282 - `"thinking-display-updates-2026-08-18"`
1283
1284 - `"ce-user-management-2026-07-13"`
1285
1286 - `"mid-conversation-output-config-2026-07-01"`
1287
1288 - `"thinking-binding-controls-2026-08-01"`
1289
1290 - `"mid-conversation-system-clear-at-2026-08-21"`
1291
1292 - `"compact-2026-09-04"`
1293
1294 - `"inline-tools-2026-09-15"`
1295
1296 - `"mcp-client-2026-09-15"`
13011297
13021298- `"anthropic-workspace-id": optional string`
13031299
from line 1926
19301926
19311927 - `string`
19321928
1933 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
1934
1935 - `"message-batches-2024-09-24"`
1936
1937 - `"prompt-caching-2024-07-31"`
1938
1939 - `"computer-use-2024-10-22"`
1940
1941 - `"computer-use-2025-01-24"`
1942
1943 - `"pdfs-2024-09-25"`
1944
1945 - `"token-counting-2024-11-01"`
1946
1947 - `"token-efficient-tools-2025-02-19"`
1948
1949 - `"output-128k-2025-02-19"`
1950
1951 - `"files-api-2025-04-14"`
1952
1953 - `"mcp-client-2025-04-04"`
1954
1955 - `"mcp-client-2025-11-20"`
1956
1957 - `"dev-full-thinking-2025-05-14"`
1958
1959 - `"interleaved-thinking-2025-05-14"`
1960
1961 - `"code-execution-2025-05-22"`
1962
1963 - `"extended-cache-ttl-2025-04-11"`
1964
1965 - `"context-1m-2025-08-07"`
1966
1967 - `"context-management-2025-06-27"`
1968
1969 - `"model-context-window-exceeded-2025-08-26"`
1970
1971 - `"skills-2025-10-02"`
1972
1973 - `"fast-mode-2026-02-01"`
1974
1975 - `"output-300k-2026-03-24"`
1976
1977 - `"user-profiles-2026-03-24"`
1978
1979 - `"user-profiles-2026-08-18"`
1980
1981 - `"user-profiles-2026-09-04"`
1982
1983 - `"advisor-tool-2026-03-01"`
1984
1985 - `"managed-agents-2026-04-01"`
1986
1987 - `"cache-diagnosis-2026-04-07"`
1988
1989 - `"dreaming-2026-04-21"`
1990
1991 - `"thinking-token-count-2026-05-13"`
1992
1993 - `"server-side-fallback-2026-06-01"`
1994
1995 - `"server-side-fallback-2026-07-01"`
1996
1997 - `"fallback-credit-2026-06-01"`
1998
1999 - `"fallback-credit-2026-07-01"`
2000
2001 - `"agent-memory-2026-07-22"`
2002
2003 - `"mid-conversation-tool-changes-2026-07-01"`
2004
2005 - `"compact-2026-01-12"`
2006
2007 - `"computer-use-2025-11-24"`
2008
2009 - `"mcp-tunnels-2026-06-22"`
2010
2011 - `"structured-outputs-2025-11-13"`
2012
2013 - `"task-budgets-2026-03-13"`
2014
2015 - `"thinking-display-updates-2026-08-18"`
2016
2017 - `"ce-user-management-2026-07-13"`
2018
2019 - `"mid-conversation-output-config-2026-07-01"`
2020
2021 - `"thinking-binding-controls-2026-08-01"`
2022
2023 - `"mid-conversation-system-clear-at-2026-08-21"`
2024
2025 - `"compact-2026-09-04"`
2026
2027 - `"inline-tools-2026-09-15"`
2028
2029 - `"mcp-client-2026-09-15"`
1929 - `"message-batches-2024-09-24"`
1930
1931 - `"prompt-caching-2024-07-31"`
1932
1933 - `"computer-use-2024-10-22"`
1934
1935 - `"computer-use-2025-01-24"`
1936
1937 - `"pdfs-2024-09-25"`
1938
1939 - `"token-counting-2024-11-01"`
1940
1941 - `"token-efficient-tools-2025-02-19"`
1942
1943 - `"output-128k-2025-02-19"`
1944
1945 - `"files-api-2025-04-14"`
1946
1947 - `"mcp-client-2025-04-04"`
1948
1949 - `"mcp-client-2025-11-20"`
1950
1951 - `"dev-full-thinking-2025-05-14"`
1952
1953 - `"interleaved-thinking-2025-05-14"`
1954
1955 - `"code-execution-2025-05-22"`
1956
1957 - `"extended-cache-ttl-2025-04-11"`
1958
1959 - `"context-1m-2025-08-07"`
1960
1961 - `"context-management-2025-06-27"`
1962
1963 - `"model-context-window-exceeded-2025-08-26"`
1964
1965 - `"skills-2025-10-02"`
1966
1967 - `"fast-mode-2026-02-01"`
1968
1969 - `"output-300k-2026-03-24"`
1970
1971 - `"user-profiles-2026-03-24"`
1972
1973 - `"user-profiles-2026-08-18"`
1974
1975 - `"user-profiles-2026-09-04"`
1976
1977 - `"advisor-tool-2026-03-01"`
1978
1979 - `"managed-agents-2026-04-01"`
1980
1981 - `"cache-diagnosis-2026-04-07"`
1982
1983 - `"dreaming-2026-04-21"`
1984
1985 - `"thinking-token-count-2026-05-13"`
1986
1987 - `"server-side-fallback-2026-06-01"`
1988
1989 - `"server-side-fallback-2026-07-01"`
1990
1991 - `"fallback-credit-2026-06-01"`
1992
1993 - `"fallback-credit-2026-07-01"`
1994
1995 - `"agent-memory-2026-07-22"`
1996
1997 - `"mid-conversation-tool-changes-2026-07-01"`
1998
1999 - `"compact-2026-01-12"`
2000
2001 - `"computer-use-2025-11-24"`
2002
2003 - `"mcp-tunnels-2026-06-22"`
2004
2005 - `"structured-outputs-2025-11-13"`
2006
2007 - `"task-budgets-2026-03-13"`
2008
2009 - `"thinking-display-updates-2026-08-18"`
2010
2011 - `"ce-user-management-2026-07-13"`
2012
2013 - `"mid-conversation-output-config-2026-07-01"`
2014
2015 - `"thinking-binding-controls-2026-08-01"`
2016
2017 - `"mid-conversation-system-clear-at-2026-08-21"`
2018
2019 - `"compact-2026-09-04"`
2020
2021 - `"inline-tools-2026-09-15"`
2022
2023 - `"mcp-client-2026-09-15"`
20302024
20312025- `"anthropic-workspace-id": optional string`
20322026
from line 2644
26502644
26512645 - `string`
26522646
2653 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
2654
2655 - `"message-batches-2024-09-24"`
2656
2657 - `"prompt-caching-2024-07-31"`
2658
2659 - `"computer-use-2024-10-22"`
2660
2661 - `"computer-use-2025-01-24"`
2662
2663 - `"pdfs-2024-09-25"`
2664
2665 - `"token-counting-2024-11-01"`
2666
2667 - `"token-efficient-tools-2025-02-19"`
2668
2669 - `"output-128k-2025-02-19"`
2670
2671 - `"files-api-2025-04-14"`
2672
2673 - `"mcp-client-2025-04-04"`
2674
2675 - `"mcp-client-2025-11-20"`
2676
2677 - `"dev-full-thinking-2025-05-14"`
2678
2679 - `"interleaved-thinking-2025-05-14"`
2680
2681 - `"code-execution-2025-05-22"`
2682
2683 - `"extended-cache-ttl-2025-04-11"`
2684
2685 - `"context-1m-2025-08-07"`
2686
2687 - `"context-management-2025-06-27"`
2688
2689 - `"model-context-window-exceeded-2025-08-26"`
2690
2691 - `"skills-2025-10-02"`
2692
2693 - `"fast-mode-2026-02-01"`
2694
2695 - `"output-300k-2026-03-24"`
2696
2697 - `"user-profiles-2026-03-24"`
2698
2699 - `"user-profiles-2026-08-18"`
2700
2701 - `"user-profiles-2026-09-04"`
2702
2703 - `"advisor-tool-2026-03-01"`
2704
2705 - `"managed-agents-2026-04-01"`
2706
2707 - `"cache-diagnosis-2026-04-07"`
2708
2709 - `"dreaming-2026-04-21"`
2710
2711 - `"thinking-token-count-2026-05-13"`
2712
2713 - `"server-side-fallback-2026-06-01"`
2714
2715 - `"server-side-fallback-2026-07-01"`
2716
2717 - `"fallback-credit-2026-06-01"`
2718
2719 - `"fallback-credit-2026-07-01"`
2720
2721 - `"agent-memory-2026-07-22"`
2722
2723 - `"mid-conversation-tool-changes-2026-07-01"`
2724
2725 - `"compact-2026-01-12"`
2726
2727 - `"computer-use-2025-11-24"`
2728
2729 - `"mcp-tunnels-2026-06-22"`
2730
2731 - `"structured-outputs-2025-11-13"`
2732
2733 - `"task-budgets-2026-03-13"`
2734
2735 - `"thinking-display-updates-2026-08-18"`
2736
2737 - `"ce-user-management-2026-07-13"`
2738
2739 - `"mid-conversation-output-config-2026-07-01"`
2740
2741 - `"thinking-binding-controls-2026-08-01"`
2742
2743 - `"mid-conversation-system-clear-at-2026-08-21"`
2744
2745 - `"compact-2026-09-04"`
2746
2747 - `"inline-tools-2026-09-15"`
2748
2749 - `"mcp-client-2026-09-15"`
2647 - `"message-batches-2024-09-24"`
2648
2649 - `"prompt-caching-2024-07-31"`
2650
2651 - `"computer-use-2024-10-22"`
2652
2653 - `"computer-use-2025-01-24"`
2654
2655 - `"pdfs-2024-09-25"`
2656
2657 - `"token-counting-2024-11-01"`
2658
2659 - `"token-efficient-tools-2025-02-19"`
2660
2661 - `"output-128k-2025-02-19"`
2662
2663 - `"files-api-2025-04-14"`
2664
2665 - `"mcp-client-2025-04-04"`
2666
2667 - `"mcp-client-2025-11-20"`
2668
2669 - `"dev-full-thinking-2025-05-14"`
2670
2671 - `"interleaved-thinking-2025-05-14"`
2672
2673 - `"code-execution-2025-05-22"`
2674
2675 - `"extended-cache-ttl-2025-04-11"`
2676
2677 - `"context-1m-2025-08-07"`
2678
2679 - `"context-management-2025-06-27"`
2680
2681 - `"model-context-window-exceeded-2025-08-26"`
2682
2683 - `"skills-2025-10-02"`
2684
2685 - `"fast-mode-2026-02-01"`
2686
2687 - `"output-300k-2026-03-24"`
2688
2689 - `"user-profiles-2026-03-24"`
2690
2691 - `"user-profiles-2026-08-18"`
2692
2693 - `"user-profiles-2026-09-04"`
2694
2695 - `"advisor-tool-2026-03-01"`
2696
2697 - `"managed-agents-2026-04-01"`
2698
2699 - `"cache-diagnosis-2026-04-07"`
2700
2701 - `"dreaming-2026-04-21"`
2702
2703 - `"thinking-token-count-2026-05-13"`
2704
2705 - `"server-side-fallback-2026-06-01"`
2706
2707 - `"server-side-fallback-2026-07-01"`
2708
2709 - `"fallback-credit-2026-06-01"`
2710
2711 - `"fallback-credit-2026-07-01"`
2712
2713 - `"agent-memory-2026-07-22"`
2714
2715 - `"mid-conversation-tool-changes-2026-07-01"`
2716
2717 - `"compact-2026-01-12"`
2718
2719 - `"computer-use-2025-11-24"`
2720
2721 - `"mcp-tunnels-2026-06-22"`
2722
2723 - `"structured-outputs-2025-11-13"`
2724
2725 - `"task-budgets-2026-03-13"`
2726
2727 - `"thinking-display-updates-2026-08-18"`
2728
2729 - `"ce-user-management-2026-07-13"`
2730
2731 - `"mid-conversation-output-config-2026-07-01"`
2732
2733 - `"thinking-binding-controls-2026-08-01"`
2734
2735 - `"mid-conversation-system-clear-at-2026-08-21"`
2736
2737 - `"compact-2026-09-04"`
2738
2739 - `"inline-tools-2026-09-15"`
2740
2741 - `"mcp-client-2026-09-15"`
27502742
27512743- `"anthropic-workspace-id": optional string`
27522744
from line 3772
37803772
37813773 - `string`
37823774
3783 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
3784
3785 - `"message-batches-2024-09-24"`
3786
3787 - `"prompt-caching-2024-07-31"`
3788
3789 - `"computer-use-2024-10-22"`
3790
3791 - `"computer-use-2025-01-24"`
3792
3793 - `"pdfs-2024-09-25"`
3794
3795 - `"token-counting-2024-11-01"`
3796
3797 - `"token-efficient-tools-2025-02-19"`
3798
3799 - `"output-128k-2025-02-19"`
3800
3801 - `"files-api-2025-04-14"`
3802
3803 - `"mcp-client-2025-04-04"`
3804
3805 - `"mcp-client-2025-11-20"`
3806
3807 - `"dev-full-thinking-2025-05-14"`
3808
3809 - `"interleaved-thinking-2025-05-14"`
3810
3811 - `"code-execution-2025-05-22"`
3812
3813 - `"extended-cache-ttl-2025-04-11"`
3814
3815 - `"context-1m-2025-08-07"`
3816
3817 - `"context-management-2025-06-27"`
3818
3819 - `"model-context-window-exceeded-2025-08-26"`
3820
3821 - `"skills-2025-10-02"`
3822
3823 - `"fast-mode-2026-02-01"`
3824
3825 - `"output-300k-2026-03-24"`
3826
3827 - `"user-profiles-2026-03-24"`
3828
3829 - `"user-profiles-2026-08-18"`
3830
3831 - `"user-profiles-2026-09-04"`
3832
3833 - `"advisor-tool-2026-03-01"`
3834
3835 - `"managed-agents-2026-04-01"`
3836
3837 - `"cache-diagnosis-2026-04-07"`
3838
3839 - `"dreaming-2026-04-21"`
3840
3841 - `"thinking-token-count-2026-05-13"`
3842
3843 - `"server-side-fallback-2026-06-01"`
3844
3845 - `"server-side-fallback-2026-07-01"`
3846
3847 - `"fallback-credit-2026-06-01"`
3848
3849 - `"fallback-credit-2026-07-01"`
3850
3851 - `"agent-memory-2026-07-22"`
3852
3853 - `"mid-conversation-tool-changes-2026-07-01"`
3854
3855 - `"compact-2026-01-12"`
3856
3857 - `"computer-use-2025-11-24"`
3858
3859 - `"mcp-tunnels-2026-06-22"`
3860
3861 - `"structured-outputs-2025-11-13"`
3862
3863 - `"task-budgets-2026-03-13"`
3864
3865 - `"thinking-display-updates-2026-08-18"`
3866
3867 - `"ce-user-management-2026-07-13"`
3868
3869 - `"mid-conversation-output-config-2026-07-01"`
3870
3871 - `"thinking-binding-controls-2026-08-01"`
3872
3873 - `"mid-conversation-system-clear-at-2026-08-21"`
3874
3875 - `"compact-2026-09-04"`
3876
3877 - `"inline-tools-2026-09-15"`
3878
3879 - `"mcp-client-2026-09-15"`
3775 - `"message-batches-2024-09-24"`
3776
3777 - `"prompt-caching-2024-07-31"`
3778
3779 - `"computer-use-2024-10-22"`
3780
3781 - `"computer-use-2025-01-24"`
3782
3783 - `"pdfs-2024-09-25"`
3784
3785 - `"token-counting-2024-11-01"`
3786
3787 - `"token-efficient-tools-2025-02-19"`
3788
3789 - `"output-128k-2025-02-19"`
3790
3791 - `"files-api-2025-04-14"`
3792
3793 - `"mcp-client-2025-04-04"`
3794
3795 - `"mcp-client-2025-11-20"`
3796
3797 - `"dev-full-thinking-2025-05-14"`
3798
3799 - `"interleaved-thinking-2025-05-14"`
3800
3801 - `"code-execution-2025-05-22"`
3802
3803 - `"extended-cache-ttl-2025-04-11"`
3804
3805 - `"context-1m-2025-08-07"`
3806
3807 - `"context-management-2025-06-27"`
3808
3809 - `"model-context-window-exceeded-2025-08-26"`
3810
3811 - `"skills-2025-10-02"`
3812
3813 - `"fast-mode-2026-02-01"`
3814
3815 - `"output-300k-2026-03-24"`
3816
3817 - `"user-profiles-2026-03-24"`
3818
3819 - `"user-profiles-2026-08-18"`
3820
3821 - `"user-profiles-2026-09-04"`
3822
3823 - `"advisor-tool-2026-03-01"`
3824
3825 - `"managed-agents-2026-04-01"`
3826
3827 - `"cache-diagnosis-2026-04-07"`
3828
3829 - `"dreaming-2026-04-21"`
3830
3831 - `"thinking-token-count-2026-05-13"`
3832
3833 - `"server-side-fallback-2026-06-01"`
3834
3835 - `"server-side-fallback-2026-07-01"`
3836
3837 - `"fallback-credit-2026-06-01"`
3838
3839 - `"fallback-credit-2026-07-01"`
3840
3841 - `"agent-memory-2026-07-22"`
3842
3843 - `"mid-conversation-tool-changes-2026-07-01"`
3844
3845 - `"compact-2026-01-12"`
3846
3847 - `"computer-use-2025-11-24"`
3848
3849 - `"mcp-tunnels-2026-06-22"`
3850
3851 - `"structured-outputs-2025-11-13"`
3852
3853 - `"task-budgets-2026-03-13"`
3854
3855 - `"thinking-display-updates-2026-08-18"`
3856
3857 - `"ce-user-management-2026-07-13"`
3858
3859 - `"mid-conversation-output-config-2026-07-01"`
3860
3861 - `"thinking-binding-controls-2026-08-01"`
3862
3863 - `"mid-conversation-system-clear-at-2026-08-21"`
3864
3865 - `"compact-2026-09-04"`
3866
3867 - `"inline-tools-2026-09-15"`
3868
3869 - `"mcp-client-2026-09-15"`
38803870
38813871- `"anthropic-workspace-id": optional string`
38823872
from line 4491
45014491
45024492 - `string`
45034493
4504 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
4505
4506 - `"message-batches-2024-09-24"`
4507
4508 - `"prompt-caching-2024-07-31"`
4509
4510 - `"computer-use-2024-10-22"`
4511
4512 - `"computer-use-2025-01-24"`
4513
4514 - `"pdfs-2024-09-25"`
4515
4516 - `"token-counting-2024-11-01"`
4517
4518 - `"token-efficient-tools-2025-02-19"`
4519
4520 - `"output-128k-2025-02-19"`
4521
4522 - `"files-api-2025-04-14"`
4523
4524 - `"mcp-client-2025-04-04"`
4525
4526 - `"mcp-client-2025-11-20"`
4527
4528 - `"dev-full-thinking-2025-05-14"`
4529
4530 - `"interleaved-thinking-2025-05-14"`
4531
4532 - `"code-execution-2025-05-22"`
4533
4534 - `"extended-cache-ttl-2025-04-11"`
4535
4536 - `"context-1m-2025-08-07"`
4537
4538 - `"context-management-2025-06-27"`
4539
4540 - `"model-context-window-exceeded-2025-08-26"`
4541
4542 - `"skills-2025-10-02"`
4543
4544 - `"fast-mode-2026-02-01"`
4545
4546 - `"output-300k-2026-03-24"`
4547
4548 - `"user-profiles-2026-03-24"`
4549
4550 - `"user-profiles-2026-08-18"`
4551
4552 - `"user-profiles-2026-09-04"`
4553
4554 - `"advisor-tool-2026-03-01"`
4555
4556 - `"managed-agents-2026-04-01"`
4557
4558 - `"cache-diagnosis-2026-04-07"`
4559
4560 - `"dreaming-2026-04-21"`
4561
4562 - `"thinking-token-count-2026-05-13"`
4563
4564 - `"server-side-fallback-2026-06-01"`
4565
4566 - `"server-side-fallback-2026-07-01"`
4567
4568 - `"fallback-credit-2026-06-01"`
4569
4570 - `"fallback-credit-2026-07-01"`
4571
4572 - `"agent-memory-2026-07-22"`
4573
4574 - `"mid-conversation-tool-changes-2026-07-01"`
4575
4576 - `"compact-2026-01-12"`
4577
4578 - `"computer-use-2025-11-24"`
4579
4580 - `"mcp-tunnels-2026-06-22"`
4581
4582 - `"structured-outputs-2025-11-13"`
4583
4584 - `"task-budgets-2026-03-13"`
4585
4586 - `"thinking-display-updates-2026-08-18"`
4587
4588 - `"ce-user-management-2026-07-13"`
4589
4590 - `"mid-conversation-output-config-2026-07-01"`
4591
4592 - `"thinking-binding-controls-2026-08-01"`
4593
4594 - `"mid-conversation-system-clear-at-2026-08-21"`
4595
4596 - `"compact-2026-09-04"`
4597
4598 - `"inline-tools-2026-09-15"`
4599
4600 - `"mcp-client-2026-09-15"`
4494 - `"message-batches-2024-09-24"`
4495
4496 - `"prompt-caching-2024-07-31"`
4497
4498 - `"computer-use-2024-10-22"`
4499
4500 - `"computer-use-2025-01-24"`
4501
4502 - `"pdfs-2024-09-25"`
4503
4504 - `"token-counting-2024-11-01"`
4505
4506 - `"token-efficient-tools-2025-02-19"`
4507
4508 - `"output-128k-2025-02-19"`
4509
4510 - `"files-api-2025-04-14"`
4511
4512 - `"mcp-client-2025-04-04"`
4513
4514 - `"mcp-client-2025-11-20"`
4515
4516 - `"dev-full-thinking-2025-05-14"`
4517
4518 - `"interleaved-thinking-2025-05-14"`
4519
4520 - `"code-execution-2025-05-22"`
4521
4522 - `"extended-cache-ttl-2025-04-11"`
4523
4524 - `"context-1m-2025-08-07"`
4525
4526 - `"context-management-2025-06-27"`
4527
4528 - `"model-context-window-exceeded-2025-08-26"`
4529
4530 - `"skills-2025-10-02"`
4531
4532 - `"fast-mode-2026-02-01"`
4533
4534 - `"output-300k-2026-03-24"`
4535
4536 - `"user-profiles-2026-03-24"`
4537
4538 - `"user-profiles-2026-08-18"`
4539
4540 - `"user-profiles-2026-09-04"`
4541
4542 - `"advisor-tool-2026-03-01"`
4543
4544 - `"managed-agents-2026-04-01"`
4545
4546 - `"cache-diagnosis-2026-04-07"`
4547
4548 - `"dreaming-2026-04-21"`
4549
4550 - `"thinking-token-count-2026-05-13"`
4551
4552 - `"server-side-fallback-2026-06-01"`
4553
4554 - `"server-side-fallback-2026-07-01"`
4555
4556 - `"fallback-credit-2026-06-01"`
4557
4558 - `"fallback-credit-2026-07-01"`
4559
4560 - `"agent-memory-2026-07-22"`
4561
4562 - `"mid-conversation-tool-changes-2026-07-01"`
4563
4564 - `"compact-2026-01-12"`
4565
4566 - `"computer-use-2025-11-24"`
4567
4568 - `"mcp-tunnels-2026-06-22"`
4569
4570 - `"structured-outputs-2025-11-13"`
4571
4572 - `"task-budgets-2026-03-13"`
4573
4574 - `"thinking-display-updates-2026-08-18"`
4575
4576 - `"ce-user-management-2026-07-13"`
4577
4578 - `"mid-conversation-output-config-2026-07-01"`
4579
4580 - `"thinking-binding-controls-2026-08-01"`
4581
4582 - `"mid-conversation-system-clear-at-2026-08-21"`
4583
4584 - `"compact-2026-09-04"`
4585
4586 - `"inline-tools-2026-09-15"`
4587
4588 - `"mcp-client-2026-09-15"`
46014589
46024590- `"anthropic-workspace-id": optional string`
46034591
from line 4872
48844872
48854873 - `string`
48864874
4887 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
4888
4889 - `"message-batches-2024-09-24"`
4890
4891 - `"prompt-caching-2024-07-31"`
4892
4893 - `"computer-use-2024-10-22"`
4894
4895 - `"computer-use-2025-01-24"`
4896
4897 - `"pdfs-2024-09-25"`
4898
4899 - `"token-counting-2024-11-01"`
4900
4901 - `"token-efficient-tools-2025-02-19"`
4902
4903 - `"output-128k-2025-02-19"`
4904
4905 - `"files-api-2025-04-14"`
4906
4907 - `"mcp-client-2025-04-04"`
4908
4909 - `"mcp-client-2025-11-20"`
4910
4911 - `"dev-full-thinking-2025-05-14"`
4912
4913 - `"interleaved-thinking-2025-05-14"`
4914
4915 - `"code-execution-2025-05-22"`
4916
4917 - `"extended-cache-ttl-2025-04-11"`
4918
4919 - `"context-1m-2025-08-07"`
4920
4921 - `"context-management-2025-06-27"`
4922
4923 - `"model-context-window-exceeded-2025-08-26"`
4924
4925 - `"skills-2025-10-02"`
4926
4927 - `"fast-mode-2026-02-01"`
4928
4929 - `"output-300k-2026-03-24"`
4930
4931 - `"user-profiles-2026-03-24"`
4932
4933 - `"user-profiles-2026-08-18"`
4934
4935 - `"user-profiles-2026-09-04"`
4936
4937 - `"advisor-tool-2026-03-01"`
4938
4939 - `"managed-agents-2026-04-01"`
4940
4941 - `"cache-diagnosis-2026-04-07"`
4942
4943 - `"dreaming-2026-04-21"`
4944
4945 - `"thinking-token-count-2026-05-13"`
4946
4947 - `"server-side-fallback-2026-06-01"`
4948
4949 - `"server-side-fallback-2026-07-01"`
4950
4951 - `"fallback-credit-2026-06-01"`
4952
4953 - `"fallback-credit-2026-07-01"`
4954
4955 - `"agent-memory-2026-07-22"`
4956
4957 - `"mid-conversation-tool-changes-2026-07-01"`
4958
4959 - `"compact-2026-01-12"`
4960
4961 - `"computer-use-2025-11-24"`
4962
4963 - `"mcp-tunnels-2026-06-22"`
4964
4965 - `"structured-outputs-2025-11-13"`
4966
4967 - `"task-budgets-2026-03-13"`
4968
4969 - `"thinking-display-updates-2026-08-18"`
4970
4971 - `"ce-user-management-2026-07-13"`
4972
4973 - `"mid-conversation-output-config-2026-07-01"`
4974
4975 - `"thinking-binding-controls-2026-08-01"`
4976
4977 - `"mid-conversation-system-clear-at-2026-08-21"`
4978
4979 - `"compact-2026-09-04"`
4980
4981 - `"inline-tools-2026-09-15"`
4982
4983 - `"mcp-client-2026-09-15"`
4875 - `"message-batches-2024-09-24"`
4876
4877 - `"prompt-caching-2024-07-31"`
4878
4879 - `"computer-use-2024-10-22"`
4880
4881 - `"computer-use-2025-01-24"`
4882
4883 - `"pdfs-2024-09-25"`
4884
4885 - `"token-counting-2024-11-01"`
4886
4887 - `"token-efficient-tools-2025-02-19"`
4888
4889 - `"output-128k-2025-02-19"`
4890
4891 - `"files-api-2025-04-14"`
4892
4893 - `"mcp-client-2025-04-04"`
4894
4895 - `"mcp-client-2025-11-20"`
4896
4897 - `"dev-full-thinking-2025-05-14"`
4898
4899 - `"interleaved-thinking-2025-05-14"`
4900
4901 - `"code-execution-2025-05-22"`
4902
4903 - `"extended-cache-ttl-2025-04-11"`
4904
4905 - `"context-1m-2025-08-07"`
4906
4907 - `"context-management-2025-06-27"`
4908
4909 - `"model-context-window-exceeded-2025-08-26"`
4910
4911 - `"skills-2025-10-02"`
4912
4913 - `"fast-mode-2026-02-01"`
4914
4915 - `"output-300k-2026-03-24"`
4916
4917 - `"user-profiles-2026-03-24"`
4918
4919 - `"user-profiles-2026-08-18"`
4920
4921 - `"user-profiles-2026-09-04"`
4922
4923 - `"advisor-tool-2026-03-01"`
4924
4925 - `"managed-agents-2026-04-01"`
4926
4927 - `"cache-diagnosis-2026-04-07"`
4928
4929 - `"dreaming-2026-04-21"`
4930
4931 - `"thinking-token-count-2026-05-13"`
4932
4933 - `"server-side-fallback-2026-06-01"`
4934
4935 - `"server-side-fallback-2026-07-01"`
4936
4937 - `"fallback-credit-2026-06-01"`
4938
4939 - `"fallback-credit-2026-07-01"`
4940
4941 - `"agent-memory-2026-07-22"`
4942
4943 - `"mid-conversation-tool-changes-2026-07-01"`
4944
4945 - `"compact-2026-01-12"`
4946
4947 - `"computer-use-2025-11-24"`
4948
4949 - `"mcp-tunnels-2026-06-22"`
4950
4951 - `"structured-outputs-2025-11-13"`
4952
4953 - `"task-budgets-2026-03-13"`
4954
4955 - `"thinking-display-updates-2026-08-18"`
4956
4957 - `"ce-user-management-2026-07-13"`
4958
4959 - `"mid-conversation-output-config-2026-07-01"`
4960
4961 - `"thinking-binding-controls-2026-08-01"`
4962
4963 - `"mid-conversation-system-clear-at-2026-08-21"`
4964
4965 - `"compact-2026-09-04"`
4966
4967 - `"inline-tools-2026-09-15"`
4968
4969 - `"mcp-client-2026-09-15"`
49844970
49854971- `"anthropic-workspace-id": optional string`
49864972
from line 5591
56055591
56065592 - `string`
56075593
5608 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
5609
5610 - `"message-batches-2024-09-24"`
5611
5612 - `"prompt-caching-2024-07-31"`
5613
5614 - `"computer-use-2024-10-22"`
5615
5616 - `"computer-use-2025-01-24"`
5617
5618 - `"pdfs-2024-09-25"`
5619
5620 - `"token-counting-2024-11-01"`
5621
5622 - `"token-efficient-tools-2025-02-19"`
5623
5624 - `"output-128k-2025-02-19"`
5625
5626 - `"files-api-2025-04-14"`
5627
5628 - `"mcp-client-2025-04-04"`
5629
5630 - `"mcp-client-2025-11-20"`
5631
5632 - `"dev-full-thinking-2025-05-14"`
5633
5634 - `"interleaved-thinking-2025-05-14"`
5635
5636 - `"code-execution-2025-05-22"`
5637
5638 - `"extended-cache-ttl-2025-04-11"`
5639
5640 - `"context-1m-2025-08-07"`
5641
5642 - `"context-management-2025-06-27"`
5643
5644 - `"model-context-window-exceeded-2025-08-26"`
5645
5646 - `"skills-2025-10-02"`
5647
5648 - `"fast-mode-2026-02-01"`
5649
5650 - `"output-300k-2026-03-24"`
5651
5652 - `"user-profiles-2026-03-24"`
5653
5654 - `"user-profiles-2026-08-18"`
5655
5656 - `"user-profiles-2026-09-04"`
5657
5658 - `"advisor-tool-2026-03-01"`
5659
5660 - `"managed-agents-2026-04-01"`
5661
5662 - `"cache-diagnosis-2026-04-07"`
5663
5664 - `"dreaming-2026-04-21"`
5665
5666 - `"thinking-token-count-2026-05-13"`
5667
5668 - `"server-side-fallback-2026-06-01"`
5669
5670 - `"server-side-fallback-2026-07-01"`
5671
5672 - `"fallback-credit-2026-06-01"`
5673
5674 - `"fallback-credit-2026-07-01"`
5675
5676 - `"agent-memory-2026-07-22"`
5677
5678 - `"mid-conversation-tool-changes-2026-07-01"`
5679
5680 - `"compact-2026-01-12"`
5681
5682 - `"computer-use-2025-11-24"`
5683
5684 - `"mcp-tunnels-2026-06-22"`
5685
5686 - `"structured-outputs-2025-11-13"`
5687
5688 - `"task-budgets-2026-03-13"`
5689
5690 - `"thinking-display-updates-2026-08-18"`
5691
5692 - `"ce-user-management-2026-07-13"`
5693
5694 - `"mid-conversation-output-config-2026-07-01"`
5695
5696 - `"thinking-binding-controls-2026-08-01"`
5697
5698 - `"mid-conversation-system-clear-at-2026-08-21"`
5699
5700 - `"compact-2026-09-04"`
5701
5702 - `"inline-tools-2026-09-15"`
5703
5704 - `"mcp-client-2026-09-15"`
5594 - `"message-batches-2024-09-24"`
5595
5596 - `"prompt-caching-2024-07-31"`
5597
5598 - `"computer-use-2024-10-22"`
5599
5600 - `"computer-use-2025-01-24"`
5601
5602 - `"pdfs-2024-09-25"`
5603
5604 - `"token-counting-2024-11-01"`
5605
5606 - `"token-efficient-tools-2025-02-19"`
5607
5608 - `"output-128k-2025-02-19"`
5609
5610 - `"files-api-2025-04-14"`
5611
5612 - `"mcp-client-2025-04-04"`
5613
5614 - `"mcp-client-2025-11-20"`
5615
5616 - `"dev-full-thinking-2025-05-14"`
5617
5618 - `"interleaved-thinking-2025-05-14"`
5619
5620 - `"code-execution-2025-05-22"`
5621
5622 - `"extended-cache-ttl-2025-04-11"`
5623
5624 - `"context-1m-2025-08-07"`
5625
5626 - `"context-management-2025-06-27"`
5627
5628 - `"model-context-window-exceeded-2025-08-26"`
5629
5630 - `"skills-2025-10-02"`
5631
5632 - `"fast-mode-2026-02-01"`
5633
5634 - `"output-300k-2026-03-24"`
5635
5636 - `"user-profiles-2026-03-24"`
5637
5638 - `"user-profiles-2026-08-18"`
5639
5640 - `"user-profiles-2026-09-04"`
5641
5642 - `"advisor-tool-2026-03-01"`
5643
5644 - `"managed-agents-2026-04-01"`
5645
5646 - `"cache-diagnosis-2026-04-07"`
5647
5648 - `"dreaming-2026-04-21"`
5649
5650 - `"thinking-token-count-2026-05-13"`
5651
5652 - `"server-side-fallback-2026-06-01"`
5653
5654 - `"server-side-fallback-2026-07-01"`
5655
5656 - `"fallback-credit-2026-06-01"`
5657
5658 - `"fallback-credit-2026-07-01"`
5659
5660 - `"agent-memory-2026-07-22"`
5661
5662 - `"mid-conversation-tool-changes-2026-07-01"`
5663
5664 - `"compact-2026-01-12"`
5665
5666 - `"computer-use-2025-11-24"`
5667
5668 - `"mcp-tunnels-2026-06-22"`
5669
5670 - `"structured-outputs-2025-11-13"`
5671
5672 - `"task-budgets-2026-03-13"`
5673
5674 - `"thinking-display-updates-2026-08-18"`
5675
5676 - `"ce-user-management-2026-07-13"`
5677
5678 - `"mid-conversation-output-config-2026-07-01"`
5679
5680 - `"thinking-binding-controls-2026-08-01"`
5681
5682 - `"mid-conversation-system-clear-at-2026-08-21"`
5683
5684 - `"compact-2026-09-04"`
5685
5686 - `"inline-tools-2026-09-15"`
5687
5688 - `"mcp-client-2026-09-15"`
57055689
57065690- `"anthropic-workspace-id": optional string`
57075691
57085692
api/beta/deployments/archive Changed · +95 / -97 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
121
122 - `"mcp-client-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
123121
124122- `"anthropic-workspace-id": optional string`
125123
api/beta/deployments/create Changed · +95 / -97 lines
from line 17
1717
1818 - `string`
1919
20 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
20 - `"message-batches-2024-09-24"`
2121
22 - `"message-batches-2024-09-24"`
22 - `"prompt-caching-2024-07-31"`
2323
24 - `"prompt-caching-2024-07-31"`
24 - `"computer-use-2024-10-22"`
2525
26 - `"computer-use-2024-10-22"`
26 - `"computer-use-2025-01-24"`
2727
28 - `"computer-use-2025-01-24"`
28 - `"pdfs-2024-09-25"`
2929
30 - `"pdfs-2024-09-25"`
30 - `"token-counting-2024-11-01"`
3131
32 - `"token-counting-2024-11-01"`
32 - `"token-efficient-tools-2025-02-19"`
3333
34 - `"token-efficient-tools-2025-02-19"`
34 - `"output-128k-2025-02-19"`
3535
36 - `"output-128k-2025-02-19"`
36 - `"files-api-2025-04-14"`
3737
38 - `"files-api-2025-04-14"`
38 - `"mcp-client-2025-04-04"`
3939
40 - `"mcp-client-2025-04-04"`
40 - `"mcp-client-2025-11-20"`
4141
42 - `"mcp-client-2025-11-20"`
42 - `"dev-full-thinking-2025-05-14"`
4343
44 - `"dev-full-thinking-2025-05-14"`
44 - `"interleaved-thinking-2025-05-14"`
4545
46 - `"interleaved-thinking-2025-05-14"`
46 - `"code-execution-2025-05-22"`
4747
48 - `"code-execution-2025-05-22"`
48 - `"extended-cache-ttl-2025-04-11"`
4949
50 - `"extended-cache-ttl-2025-04-11"`
50 - `"context-1m-2025-08-07"`
5151
52 - `"context-1m-2025-08-07"`
52 - `"context-management-2025-06-27"`
5353
54 - `"context-management-2025-06-27"`
54 - `"model-context-window-exceeded-2025-08-26"`
5555
56 - `"model-context-window-exceeded-2025-08-26"`
56 - `"skills-2025-10-02"`
5757
58 - `"skills-2025-10-02"`
58 - `"fast-mode-2026-02-01"`
5959
60 - `"fast-mode-2026-02-01"`
60 - `"output-300k-2026-03-24"`
6161
62 - `"output-300k-2026-03-24"`
62 - `"user-profiles-2026-03-24"`
6363
64 - `"user-profiles-2026-03-24"`
64 - `"user-profiles-2026-08-18"`
6565
66 - `"user-profiles-2026-08-18"`
66 - `"user-profiles-2026-09-04"`
6767
68 - `"user-profiles-2026-09-04"`
68 - `"advisor-tool-2026-03-01"`
6969
70 - `"advisor-tool-2026-03-01"`
70 - `"managed-agents-2026-04-01"`
7171
72 - `"managed-agents-2026-04-01"`
72 - `"cache-diagnosis-2026-04-07"`
7373
74 - `"cache-diagnosis-2026-04-07"`
74 - `"dreaming-2026-04-21"`
7575
76 - `"dreaming-2026-04-21"`
76 - `"thinking-token-count-2026-05-13"`
7777
78 - `"thinking-token-count-2026-05-13"`
78 - `"server-side-fallback-2026-06-01"`
7979
80 - `"server-side-fallback-2026-06-01"`
80 - `"server-side-fallback-2026-07-01"`
8181
82 - `"server-side-fallback-2026-07-01"`
82 - `"fallback-credit-2026-06-01"`
8383
84 - `"fallback-credit-2026-06-01"`
84 - `"fallback-credit-2026-07-01"`
8585
86 - `"fallback-credit-2026-07-01"`
86 - `"agent-memory-2026-07-22"`
8787
88 - `"agent-memory-2026-07-22"`
88 - `"mid-conversation-tool-changes-2026-07-01"`
8989
90 - `"mid-conversation-tool-changes-2026-07-01"`
90 - `"compact-2026-01-12"`
9191
92 - `"compact-2026-01-12"`
92 - `"computer-use-2025-11-24"`
9393
94 - `"computer-use-2025-11-24"`
94 - `"mcp-tunnels-2026-06-22"`
9595
96 - `"mcp-tunnels-2026-06-22"`
96 - `"structured-outputs-2025-11-13"`
9797
98 - `"structured-outputs-2025-11-13"`
98 - `"task-budgets-2026-03-13"`
9999
100 - `"task-budgets-2026-03-13"`
100 - `"thinking-display-updates-2026-08-18"`
101101
102 - `"thinking-display-updates-2026-08-18"`
102 - `"ce-user-management-2026-07-13"`
103103
104 - `"ce-user-management-2026-07-13"`
104 - `"mid-conversation-output-config-2026-07-01"`
105105
106 - `"mid-conversation-output-config-2026-07-01"`
106 - `"thinking-binding-controls-2026-08-01"`
107107
108 - `"thinking-binding-controls-2026-08-01"`
108 - `"mid-conversation-system-clear-at-2026-08-21"`
109109
110 - `"mid-conversation-system-clear-at-2026-08-21"`
110 - `"compact-2026-09-04"`
111111
112 - `"compact-2026-09-04"`
112 - `"inline-tools-2026-09-15"`
113113
114 - `"inline-tools-2026-09-15"`
115
116 - `"mcp-client-2026-09-15"`
114 - `"mcp-client-2026-09-15"`
117115
118116- `"anthropic-workspace-id": optional string`
119117
api/beta/deployments/list Changed · +95 / -97 lines
from line 61
6161
6262 - `string`
6363
64 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
64 - `"message-batches-2024-09-24"`
6565
66 - `"message-batches-2024-09-24"`
66 - `"prompt-caching-2024-07-31"`
6767
68 - `"prompt-caching-2024-07-31"`
68 - `"computer-use-2024-10-22"`
6969
70 - `"computer-use-2024-10-22"`
70 - `"computer-use-2025-01-24"`
7171
72 - `"computer-use-2025-01-24"`
72 - `"pdfs-2024-09-25"`
7373
74 - `"pdfs-2024-09-25"`
74 - `"token-counting-2024-11-01"`
7575
76 - `"token-counting-2024-11-01"`
76 - `"token-efficient-tools-2025-02-19"`
7777
78 - `"token-efficient-tools-2025-02-19"`
78 - `"output-128k-2025-02-19"`
7979
80 - `"output-128k-2025-02-19"`
80 - `"files-api-2025-04-14"`
8181
82 - `"files-api-2025-04-14"`
82 - `"mcp-client-2025-04-04"`
8383
84 - `"mcp-client-2025-04-04"`
84 - `"mcp-client-2025-11-20"`
8585
86 - `"mcp-client-2025-11-20"`
86 - `"dev-full-thinking-2025-05-14"`
8787
88 - `"dev-full-thinking-2025-05-14"`
88 - `"interleaved-thinking-2025-05-14"`
8989
90 - `"interleaved-thinking-2025-05-14"`
90 - `"code-execution-2025-05-22"`
9191
92 - `"code-execution-2025-05-22"`
92 - `"extended-cache-ttl-2025-04-11"`
9393
94 - `"extended-cache-ttl-2025-04-11"`
94 - `"context-1m-2025-08-07"`
9595
96 - `"context-1m-2025-08-07"`
96 - `"context-management-2025-06-27"`
9797
98 - `"context-management-2025-06-27"`
98 - `"model-context-window-exceeded-2025-08-26"`
9999
100 - `"model-context-window-exceeded-2025-08-26"`
100 - `"skills-2025-10-02"`
101101
102 - `"skills-2025-10-02"`
102 - `"fast-mode-2026-02-01"`
103103
104 - `"fast-mode-2026-02-01"`
104 - `"output-300k-2026-03-24"`
105105
106 - `"output-300k-2026-03-24"`
106 - `"user-profiles-2026-03-24"`
107107
108 - `"user-profiles-2026-03-24"`
108 - `"user-profiles-2026-08-18"`
109109
110 - `"user-profiles-2026-08-18"`
110 - `"user-profiles-2026-09-04"`
111111
112 - `"user-profiles-2026-09-04"`
112 - `"advisor-tool-2026-03-01"`
113113
114 - `"advisor-tool-2026-03-01"`
114 - `"managed-agents-2026-04-01"`
115115
116 - `"managed-agents-2026-04-01"`
116 - `"cache-diagnosis-2026-04-07"`
117117
118 - `"cache-diagnosis-2026-04-07"`
118 - `"dreaming-2026-04-21"`
119119
120 - `"dreaming-2026-04-21"`
120 - `"thinking-token-count-2026-05-13"`
121121
122 - `"thinking-token-count-2026-05-13"`
122 - `"server-side-fallback-2026-06-01"`
123123
124 - `"server-side-fallback-2026-06-01"`
124 - `"server-side-fallback-2026-07-01"`
125125
126 - `"server-side-fallback-2026-07-01"`
126 - `"fallback-credit-2026-06-01"`
127127
128 - `"fallback-credit-2026-06-01"`
128 - `"fallback-credit-2026-07-01"`
129129
130 - `"fallback-credit-2026-07-01"`
130 - `"agent-memory-2026-07-22"`
131131
132 - `"agent-memory-2026-07-22"`
132 - `"mid-conversation-tool-changes-2026-07-01"`
133133
134 - `"mid-conversation-tool-changes-2026-07-01"`
134 - `"compact-2026-01-12"`
135135
136 - `"compact-2026-01-12"`
136 - `"computer-use-2025-11-24"`
137137
138 - `"computer-use-2025-11-24"`
138 - `"mcp-tunnels-2026-06-22"`
139139
140 - `"mcp-tunnels-2026-06-22"`
140 - `"structured-outputs-2025-11-13"`
141141
142 - `"structured-outputs-2025-11-13"`
142 - `"task-budgets-2026-03-13"`
143143
144 - `"task-budgets-2026-03-13"`
144 - `"thinking-display-updates-2026-08-18"`
145145
146 - `"thinking-display-updates-2026-08-18"`
146 - `"ce-user-management-2026-07-13"`
147147
148 - `"ce-user-management-2026-07-13"`
148 - `"mid-conversation-output-config-2026-07-01"`
149149
150 - `"mid-conversation-output-config-2026-07-01"`
150 - `"thinking-binding-controls-2026-08-01"`
151151
152 - `"thinking-binding-controls-2026-08-01"`
152 - `"mid-conversation-system-clear-at-2026-08-21"`
153153
154 - `"mid-conversation-system-clear-at-2026-08-21"`
154 - `"compact-2026-09-04"`
155155
156 - `"compact-2026-09-04"`
156 - `"inline-tools-2026-09-15"`
157157
158 - `"inline-tools-2026-09-15"`
159
160 - `"mcp-client-2026-09-15"`
158 - `"mcp-client-2026-09-15"`
161159
162160- `"anthropic-workspace-id": optional string`
163161
api/beta/deployments/pause Changed · +95 / -97 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
121
122 - `"mcp-client-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
123121
124122- `"anthropic-workspace-id": optional string`
125123
api/beta/deployments/retrieve Changed · +95 / -97 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
121
122 - `"mcp-client-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
123121
124122- `"anthropic-workspace-id": optional string`
125123
api/beta/deployments/run Changed · +95 / -97 lines
from line 23
2323
2424 - `string`
2525
26 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 45 more`
26 - `"message-batches-2024-09-24"`
2727
28 - `"message-batches-2024-09-24"`
28 - `"prompt-caching-2024-07-31"`
2929
30 - `"prompt-caching-2024-07-31"`
30 - `"computer-use-2024-10-22"`
3131
32 - `"computer-use-2024-10-22"`
32 - `"computer-use-2025-01-24"`
3333
34 - `"computer-use-2025-01-24"`
34 - `"pdfs-2024-09-25"`
3535
36 - `"pdfs-2024-09-25"`
36 - `"token-counting-2024-11-01"`
3737
38 - `"token-counting-2024-11-01"`
38 - `"token-efficient-tools-2025-02-19"`
3939
40 - `"token-efficient-tools-2025-02-19"`
40 - `"output-128k-2025-02-19"`
4141
42 - `"output-128k-2025-02-19"`
42 - `"files-api-2025-04-14"`
4343
44 - `"files-api-2025-04-14"`
44 - `"mcp-client-2025-04-04"`
4545
46 - `"mcp-client-2025-04-04"`
46 - `"mcp-client-2025-11-20"`
4747
48 - `"mcp-client-2025-11-20"`
48 - `"dev-full-thinking-2025-05-14"`
4949
50 - `"dev-full-thinking-2025-05-14"`
50 - `"interleaved-thinking-2025-05-14"`
5151
52 - `"interleaved-thinking-2025-05-14"`
52 - `"code-execution-2025-05-22"`
5353
54 - `"code-execution-2025-05-22"`
54 - `"extended-cache-ttl-2025-04-11"`
5555
56 - `"extended-cache-ttl-2025-04-11"`
56 - `"context-1m-2025-08-07"`
5757
58 - `"context-1m-2025-08-07"`
58 - `"context-management-2025-06-27"`
5959
60 - `"context-management-2025-06-27"`
60 - `"model-context-window-exceeded-2025-08-26"`
6161
62 - `"model-context-window-exceeded-2025-08-26"`
62 - `"skills-2025-10-02"`
6363
64 - `"skills-2025-10-02"`
64 - `"fast-mode-2026-02-01"`
6565
66 - `"fast-mode-2026-02-01"`
66 - `"output-300k-2026-03-24"`
6767
68 - `"output-300k-2026-03-24"`
68 - `"user-profiles-2026-03-24"`
6969
70 - `"user-profiles-2026-03-24"`
70 - `"user-profiles-2026-08-18"`
7171
72 - `"user-profiles-2026-08-18"`
72 - `"user-profiles-2026-09-04"`
7373
74 - `"user-profiles-2026-09-04"`
74 - `"advisor-tool-2026-03-01"`
7575
76 - `"advisor-tool-2026-03-01"`
76 - `"managed-agents-2026-04-01"`
7777
78 - `"managed-agents-2026-04-01"`
78 - `"cache-diagnosis-2026-04-07"`
7979
80 - `"cache-diagnosis-2026-04-07"`
80 - `"dreaming-2026-04-21"`
8181
82 - `"dreaming-2026-04-21"`
82 - `"thinking-token-count-2026-05-13"`
8383
84 - `"thinking-token-count-2026-05-13"`
84 - `"server-side-fallback-2026-06-01"`
8585
86 - `"server-side-fallback-2026-06-01"`
86 - `"server-side-fallback-2026-07-01"`
8787
88 - `"server-side-fallback-2026-07-01"`
88 - `"fallback-credit-2026-06-01"`
8989
90 - `"fallback-credit-2026-06-01"`
90 - `"fallback-credit-2026-07-01"`
9191
92 - `"fallback-credit-2026-07-01"`
92 - `"agent-memory-2026-07-22"`
9393
94 - `"agent-memory-2026-07-22"`
94 - `"mid-conversation-tool-changes-2026-07-01"`
9595
96 - `"mid-conversation-tool-changes-2026-07-01"`
96 - `"compact-2026-01-12"`
9797
98 - `"compact-2026-01-12"`
98 - `"computer-use-2025-11-24"`
9999
100 - `"computer-use-2025-11-24"`
100 - `"mcp-tunnels-2026-06-22"`
101101
102 - `"mcp-tunnels-2026-06-22"`
102 - `"structured-outputs-2025-11-13"`
103103
104 - `"structured-outputs-2025-11-13"`
104 - `"task-budgets-2026-03-13"`
105105
106 - `"task-budgets-2026-03-13"`
106 - `"thinking-display-updates-2026-08-18"`
107107
108 - `"thinking-display-updates-2026-08-18"`
108 - `"ce-user-management-2026-07-13"`
109109
110 - `"ce-user-management-2026-07-13"`
110 - `"mid-conversation-output-config-2026-07-01"`
111111
112 - `"mid-conversation-output-config-2026-07-01"`
112 - `"thinking-binding-controls-2026-08-01"`
113113
114 - `"thinking-binding-controls-2026-08-01"`
114 - `"mid-conversation-system-clear-at-2026-08-21"`
115115
116 - `"mid-conversation-system-clear-at-2026-08-21"`
116 - `"compact-2026-09-04"`
117117
118 - `"compact-2026-09-04"`
118 - `"inline-tools-2026-09-15"`
119119
120 - `"inline-tools-2026-09-15"`
121
122 - `"mcp-client-2026-09-15"`
120 - `"mcp-client-2026-09-15"`
123121
124122- `"anthropic-workspace-id": optional string`
125123