One read of Claude Developer Platformapi-20261016T013710Z
154 pages moved out of 782 read.
Pages moved
154
significant first
Pages read
782
in this capture
Captured
01:37 UTC
Corpus hash
c85f02ea566f
corpus-hash
What this read moved
126-150 of 154, page 6 of 7This capture is too large to show at once. Changes 126-150 of 154 are below, significant first; the rest are on the following screens.
api/compliance/apps/sessions/local Changed · +18 / -6 lines
from line 18
1818
1919- `created_at: optional object`
2020
21 - `gt: optional string`
22
23 Only return sessions whose first inference call is strictly after this time (RFC 3339; a UTC offset is required).
24
25 format: date-time
26
2127 - `gte: optional string`
2228
2329 Only return sessions whose first inference call is at or after this time (RFC 3339; a UTC offset is required).
from line 36
3036
3137 format: date-time
3238
39 - `lte: optional string`
40
41 Only return sessions whose first inference call is at or before this time (RFC 3339; a UTC offset is required).
42
43 format: date-time
44
3345- `limit: optional number`
3446
3547 Maximum results (default: 100, max: 500)
from line 56
4456
4557 - `gte: optional string`
4658
47 Only return sessions whose last inference call is at or after this time (RFC 3339; a UTC offset is required). Combines with `created_at.gte` / `created_at.lt`; the ordering and pagination are unchanged. Use it to poll for sessions that have been active since a previous pass — a session that becomes active later can only enter the result, never leave it.
59 Only return sessions whose last inference call is at or after this time (RFC 3339; a UTC offset is required). Combines with the `created_at` filters and does not change the ordering or pagination. Use it to poll for sessions that have been active since a previous pass — a session that becomes active later can only enter the result, never leave it.
4860
4961 format: date-time
5062
from line 106
94106
95107 - `updated_at: string`
96108
97 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
109 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
98110
99111 format: date-time
100112
from line 217
205217
206218- `updated_at: string`
207219
208 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
220 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
209221
210222 format: date-time
211223
from line 302
290302
291303 - `updated_at: string`
292304
293 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
305 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
294306
295307 format: date-time
296308
from line 359
347359
348360 - `updated_at: string`
349361
350 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
362 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
351363
352364 format: date-time
353365
from line 652
640652
641653 - `updated_at: string`
642654
643 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
655 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
644656
645657 format: date-time
646658
api/compliance/apps/sessions/local/list Changed · +14 / -2 lines
from line 16
1616
1717- `created_at: optional object`
1818
19 - `gt: optional string`
20
21 Only return sessions whose first inference call is strictly after this time (RFC 3339; a UTC offset is required).
22
23 format: date-time
24
1925 - `gte: optional string`
2026
2127 Only return sessions whose first inference call is at or after this time (RFC 3339; a UTC offset is required).
from line 34
2834
2935 format: date-time
3036
37 - `lte: optional string`
38
39 Only return sessions whose first inference call is at or before this time (RFC 3339; a UTC offset is required).
40
41 format: date-time
42
3143- `limit: optional number`
3244
3345 Maximum results (default: 100, max: 500)
from line 54
4254
4355 - `gte: optional string`
4456
45 Only return sessions whose last inference call is at or after this time (RFC 3339; a UTC offset is required). Combines with `created_at.gte` / `created_at.lt`; the ordering and pagination are unchanged. Use it to poll for sessions that have been active since a previous pass — a session that becomes active later can only enter the result, never leave it.
57 Only return sessions whose last inference call is at or after this time (RFC 3339; a UTC offset is required). Combines with the `created_at` filters and does not change the ordering or pagination. Use it to poll for sessions that have been active since a previous pass — a session that becomes active later can only enter the result, never leave it.
4658
4759 format: date-time
4860
from line 104
92104
93105 - `updated_at: string`
94106
95 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
107 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
96108
97109 format: date-time
98110
api/messages Changed · +80 / -1 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.
from line 1275
12751275
12761276 - `"standard_only"`
12771277
1278- `speed: optional "standard" or "fast" or null`
1279
1280 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
1281
1282 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
1283
1284 - `"standard"`
1285
1286 - `"fast"`
1287
12781288- `stop_sequences: optional array of string`
12791289
12801290 Custom text sequences that will cause the model to stop generating.
from line 4210
42004210
42014211 - `"batch"`
42024212
4213 - `speed: "standard" or "fast" or null`
4214
4215 The inference speed mode used for this request.
4216
4217 - `"standard"`
4218
4219 - `"fast"`
4220
42034221- `RawMessageStreamEvent = RawMessageStartEvent or RawMessageDeltaEvent or RawMessageStopEvent or 3 more`
42044222
42054223 - `RawMessageStartEvent object`
from line 4527
45094527 "web_fetch_requests": 2,
45104528 "web_search_requests": 0
45114529 },
4512 "service_tier": "standard"
4530 "service_tier": "standard",
4531 "speed": "standard"
45134532 }
45144533}
45154534```
from line 5715
56965715
56975716 The JSON schema of the format
56985717
5718- `speed: optional "standard" or "fast" or null`
5719
5720 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
5721
5722 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
5723
5724 - `"standard"`
5725
5726 - `"fast"`
5727
56995728- `system: optional string or array of TextBlockParam`
57005729
57015730 System prompt.
from line 8559
85308559
85318560 default: tool_use
85328561
8533 - `id: string`
8534
8535 pattern: ^[a-zA-Z0-9_-]+$
8536
8537 - `caller: ToolUseCaller`
8538
8539 Which party invoked the tool call: the model directly, or a server tool on its behalf.
8540
8541 - `DirectCaller object`
8542
8543 Tool invocation directly from the model.
8544
8545 - `type: "direct"`
8546
8547 - `ServerToolCaller object`
8548
8549 Tool invocation generated by a server-side tool.
8550
8551 - `type: "code_execution_20250825"`
8552
8553 - `tool_id: string`
8554
8555 pattern: ^srvtoolu_[a-zA-Z0-9_]+$
8556
8557 - `ServerToolCaller20260120 object`
8558
8559 - `type: "code_execution_20260120"`
8560
8561 - `tool_id: string`
8562
8563 pattern: ^srvtoolu_[a-zA-Z0-9_]+$
8564
8565 - `input: BrowserHoldKeyInput`
8566
8567 Hold a key or key chord down for a duration, then release it. Uses the same key
8568 names and "+" chord syntax as the key action.
8569
8570 - `duration: number`
8571
8572 Seconds to hold the key down (maximum 30).
8573
8574 minimum: 0, maximum: 30
8575
8576 - `text: string`
8577
8578 The key or chord
8562 - `i
api/messages/batches Changed · +42 / -0 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.
from line 1295
12951295
12961296 - `"standard_only"`
12971297
1298 - `speed: optional "standard" or "fast" or null`
1299
1300 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
1301
1302 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
1303
1304 - `"standard"`
1305
1306 - `"fast"`
1307
12981308 - `stop_sequences: optional array of string`
12991309
13001310 Custom text sequences that will cause the model to stop generating.
from line 5046
50365046
50375047 - `"batch"`
50385048
5049 - `speed: "standard" or "fast" or null`
5050
5051 The inference speed mode used for this request.
5052
5053 - `"standard"`
5054
5055 - `"fast"`
5056
50395057 - `MessageBatchErroredResult object`
50405058
50415059 - `type: "errored"`
from line 6567
65496567
65506568 - `"batch"`
65516569
6570 - `speed: "standard" or "fast" or null`
6571
6572 The inference speed mode used for this request.
6573
6574 - `"standard"`
6575
6576 - `"fast"`
6577
65526578 - `MessageBatchErroredResult object`
65536579
65546580 - `type: "errored"`
from line 7860
78347860
78357861 - `"batch"`
78367862
7863 - `speed: "standard" or "fast" or null`
7864
7865 The inference speed mode used for this request.
7866
7867 - `"standard"`
7868
7869 - `"fast"`
7870
78377871 - `MessageBatchErroredResult object`
78387872
78397873 - `type: "errored"`
from line 8641
86078641
86088642 - `TextEditorCodeExecutionToolResultError object`
86098643
8610 - `type: "text_editor_code_execution_tool_result_error"`
8611
8612 default: text_editor_code_execution_tool_result_error
8613
8614 - `error_code: TextEditorCodeExecutionToolResultErrorCode`
8615
8616 - `"invalid_tool_input"`
8617
8618 - `"unavailable"`
8619
8620 - `"too_many_requests"`
8621
8622 - `"execution_time_exceeded"`
8623
8624 - `"file_not_found"`
8625
8626 - `error_message: string or null`
8627
8628 - `TextEditorCodeExecutionViewResultBlock object`
8629
8630 - `type: "text_editor_code_execution_view_result"`
8631
8632 default: text_editor_code_execution_view_result
8633
8634 - `content: string`
8635
8636 - `file_type: "text" or "image" or "pdf"`
8637
8638 - `"text"`
8639
8640 - `"image"`
8641
8642 - `"pdf"`
8643
8644 - `num_lines: number or null`
8645
8646 - `start_line: number or null`
8647
8648 - `total_lines: n
8644 - `type: "text_e
api/messages/batches/create Changed · +10 / -0 lines
from line 1293
12931293
12941294 - `"standard_only"`
12951295
1296 - `speed: optional "standard" or "fast" or null`
1297
1298 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
1299
1300 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
1301
1302 - `"standard"`
1303
1304 - `"fast"`
1305
12961306 - `stop_sequences: optional array of string`
12971307
12981308 Custom text sequences that will cause the model to stop generating.
api/messages/batches/results Changed · +8 / -0 lines
from line 1168
11681168
11691169 - `"batch"`
11701170
1171 - `speed: "standard" or "fast" or null`
1172
1173 The inference speed mode used for this request.
1174
1175 - `"standard"`
1176
1177 - `"fast"`
1178
11711179 - `MessageBatchErroredResult object`
11721180
11731181 - `type: "errored"`
api/messages/count_tokens Changed · +10 / -0 lines
from line 1185
11851185
11861186 The JSON schema of the format
11871187
1188- `speed: optional "standard" or "fast" or null`
1189
1190 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
1191
1192 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
1193
1194 - `"standard"`
1195
1196 - `"fast"`
1197
11881198- `system: optional string or array of TextBlockParam`
11891199
11901200 System prompt.
api/messages/create Changed · +20 / -1 lines
from line 1273
12731273
12741274 - `"standard_only"`
12751275
1276- `speed: optional "standard" or "fast" or null`
1277
1278 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
1279
1280 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
1281
1282 - `"standard"`
1283
1284 - `"fast"`
1285
12761286- `stop_sequences: optional array of string`
12771287
12781288 Custom text sequences that will cause the model to stop generating.
from line 4208
41984208
41994209 - `"batch"`
42004210
4211 - `speed: "standard" or "fast" or null`
4212
4213 The inference speed mode used for this request.
4214
4215 - `"standard"`
4216
4217 - `"fast"`
4218
42014219- `RawMessageStreamEvent = RawMessageStartEvent or RawMessageDeltaEvent or RawMessageStopEvent or 3 more`
42024220
42034221 - `RawMessageStartEvent object`
from line 4525
45074525 "web_fetch_requests": 2,
45084526 "web_search_requests": 0
45094527 },
4510 "service_tier": "standard"
4528 "service_tier": "standard",
4529 "speed": "standard"
45114530 }
45124531}
45134532```
build-with-claude/fast-mode Changed · +16 / -27 lines
## Supported models and platforms ## Supported models
from line 14
1414 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
1515</Note>
1616
17## Supported models
17## Supported models and platforms
1818
19Fast mode is supported on the following models:
19Fast mode is supported on the following models and platforms:
2020
21* Claude Opus 5.5 (claude-opus-5-5)
22* Claude Opus 5 (claude-opus-5)
23* Claude Opus 4.8 (claude-opus-4-8)
21| Supported model | Supported platforms |
22| --------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
23| Claude Opus 5.5 (claude-opus-5-5) | Claude API, including [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)<br />[Claude in Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry), for deployments hosted on Azure |
24| Claude Opus 5 (claude-opus-5) | Claude API, including [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) |
25| Claude Opus 4.8 (claude-opus-4-8) | Claude API, including [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) |
2426
25<Note>
26 Fast mode is available on the Claude API, including [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview). Fast mode for Claude Opus 5.5 is also available on [Claude in Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) for deployments hosted on Azure, but not for deployments hosted on Anthropic. Fast mode is not available on Amazon Bedrock, Claude Platform on AWS, or Google Cloud.
27</Note>
27If you were using fast mode on an earlier model, [migrate to a model that supports fast mode](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide).
2828
29<Note>
30 Fast mode is not available on Claude Opus 4.7. Requests to `claude-opus-4-7` with `speed: "fast"` return an error; unlike Claude Opus 4.6 (see the following note), requests do not fall back to standard speed. The model itself remains available at standard speed. To continue using fast mode, see [Migrating to Claude Opus 5.5 from Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-47).
31</Note>
29Fast mode is not available for Claude in Microsoft Foundry deployments [hosted on Anthropic](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry#hosting-options).
3230
33<Note>
34 Fast mode is not available on Claude Opus 4.6. Requests to `claude-opus-4-6` that set `speed` need the `fast-mode-2026-02-01` beta header; without it, they return a `400` error. With the header, requests with `speed: "fast"` run at standard speed and are billed at [standard rates](https://platform.claude.com/docs/en/about-claude/pricing) rather than fast mode's premium rates, and the response reports [`usage.speed: "standard"`](https://platform.claude.com/docs/en/build-with-claude/fast-mode#checking-which-speed-was-used). To continue using fast mode, see [Migrating to Claude Opus 5.5 from Claude Opus 4.6 and earlier Opus models](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-46).
35</Note>
36
3731## How fast mode works
3832
3933Fast mode runs the same model with a faster inference configuration. There is no change to intelligence or capabilities.
from line 179
185179 );
186180
187181 foreach ($response->content as $block) {
188 if ($block->type === 'text') {
182 if ($block instanceof \Anthropic\Messages\TextBlock) {
189183 echo $block->text, PHP_EOL;
190184 }
191185 }
from line 221
227221
228222Fast mode has a dedicated rate limit that is separate from standard Opus rate limits. When your fast mode rate limit is exceeded, the API returns a `429` error with a `retry-after` header indicating when capacity will be available.
229223
230Claude Opus 5.5 has a default fast mode rate limit in every [usage tier](https://platform.claude.com/docs/en/api/rate-limits), so you don't need to request access to use fast mode with it.
224Claude Opus 5.5 has a default fast mode rate limit in the Start, Build, and Scale [usage tiers](https://platform.claude.com/docs/en/api/rate-limits), so you don't need to request access to use fast mode with it. Organizations in the Evaluation tier can use fast mode with Claude Opus 5.5 once they move to the Start, Build, or Scale tier.
231225
232226The response includes headers that indicate your fast mode rate limit status:
233227
from line 238
244238
245239## Checking which speed was used
246240
247The response `usage` object includes a `speed` field that indicates which speed was used, either `"fast"` or `"standard"`. Requesting `speed: "fast"` on a [model that doesn't support fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models) returns an error, and so does exceeding fast mode's rate limits or capacity (a `429` or `529`). When a request with `speed: "fast"` succeeds, `usage.speed` is `"fast"`. Claude Opus 4.6 is an exception: a request with `speed: "fast"` and the `fast-mode-2026-02-01` beta header does not return an error but runs at standard speed, and `usage.speed` shows `"standard"`. Without the header, the request returns a `400` error.
241The response `usage` object includes a `speed` field that indicates which speed was used, either `"fast"` or `"standard"`. Requesting `speed: "fast"` on a [model that doesn't support fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models-and-platforms) returns an error, and so does exceeding fast mode's rate limits or capacity (a `429` or `529`). When a request with `speed: "fast"` succeeds, `usage.speed` is `"fast"`. Claude Opus 4.6 is an exception: instead of returning an error, a request with `speed: "fast"` and the `fast-mode-2026-02-01` beta header succeeds, but runs at standard speed. This behavior is correctly reflected in `usage.speed`, which shows `"standard"` in this situation.
248242
249243<CodeGroup>
250244 ```bash cURL
from line 389
395389
396390### Falling back to standard speed
397391
398<Note>
399 This section covers an opt-in client-side fallback when fast mode is rate limited. It is separate from the behavior on [Claude Opus 4.6](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models), where fast mode is not available.
400</Note>
401
402392If you'd prefer to fall back to standard speed rather than wait for fast mode capacity, catch the rate limit error and retry without `speed: "fast"`. Set `max_retries` (typescript, java, php: `maxRetries`; csharp: `MaxRetries`; go: `option.WithMaxRetries`) to `0` on the initial fast request to skip automatic retries and fail immediately on rate limit errors.
403393
404394<Note>
from line 721
731721## Considerations
732722
733723* **Prompt caching:** Switching between fast and standard speed invalidates the prompt cache. Requests at different speeds do not share cached prefixes.
734* **Supported models:** Fast mode is supported on Claude Opus 5.5, Claude Opus 5, and Claude Opus 4.8. See [Supported models](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models).
735724* **TTFT:** Fast mode's benefits are focused on output tokens per second (OTPS), not time to first token (TTFT).
736725* **Batch API:** Fast mode is not available with the [Batch API](https://platform.claude.com/docs/en/build-with-claude/batch-processing).
737726* **Priority Tier:** Fast mode is not available with a [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers) commitment.
manage-claude/cmek Changed · +4 / -2 lines
from line 4
44description: Encrypt Claude workspace data at rest with a key you control.
55---
66
7```bash Learn more with the /claude-api skill in Claude Code
8claude "/claude-api tell me about customer-managed encryption keys"
7Ask Claude Code about this feature
8
9```text
10/claude-api tell me about customer-managed encryption keys
911```
1012
1113A customer-managed encryption key (CMEK) lets you provision an encryption key in your own [AWS KMS](https://aws.amazon.com/kms/), [Google Cloud KMS](https://cloud.google.com/security/products/security-key-management), or [Azure Key Vault](https://azure.microsoft.com/en-us/products/key-vault) and have Anthropic use it to encrypt certain workspace data at rest. You retain full control of the key, including rotation, audit, and revocation, and the key operations Anthropic performs against your key are recorded in your cloud provider's audit logs.
manage-claude/cmek-aws-kms Changed · +4 / -2 lines
from line 4
44description: Use AWS KMS to provide an encryption key for your organization.
55---
66
7```bash Configure with the /claude-api skill in Claude Code
8claude "/claude-api help me configure a customer-managed encryption key with AWS KMS"
7Let Claude Code set up this feature for you
8
9```text
10/claude-api help me configure a customer-managed encryption key with AWS KMS
911```
1012
1113This guide walks through configuring an [AWS KMS](https://aws.amazon.com/kms/) key as a [customer-managed encryption key (CMEK)](https://platform.claude.com/docs/en/manage-claude/cmek) for your Anthropic organization.
manage-claude/cmek-azure-key-vault Changed · +4 / -2 lines
from line 4
44description: Use Azure Key Vault to provide an encryption key for your organization.
55---
66
7```bash Configure with the /claude-api skill in Claude Code
8claude "/claude-api help me configure a customer-managed encryption key with Azure Key Vault"
7Let Claude Code set up this feature for you
8
9```text
10/claude-api help me configure a customer-managed encryption key with Azure Key Vault
911```
1012
1113This guide walks through configuring an Azure Key Vault key as a [customer-managed encryption key (CMEK)](https://platform.claude.com/docs/en/manage-claude/cmek) for your Anthropic organization.
manage-claude/cmek-google-cloud-kms Changed · +4 / -2 lines
from line 4
44description: Use Google Cloud KMS to provide an encryption key for your organization.
55---
66
7```bash Configure with the /claude-api skill in Claude Code
8claude "/claude-api help me configure a customer-managed encryption key with Google Cloud KMS"
7Let Claude Code set up this feature for you
8
9```text
10/claude-api help me configure a customer-managed encryption key with Google Cloud KMS
911```
1012
1113This guide walks through configuring a Google Cloud KMS key as a [customer-managed encryption key (CMEK)](https://platform.claude.com/docs/en/manage-claude/cmek) for your Anthropic organization.
release-notes/overview Changed · +1 / -7 lines
from line 15
1515### October 15, 2026
1616
1717* The [Compliance API](https://platform.claude.com/docs/en/manage-claude/compliance-api) now returns the text each organization member saved as **Instructions for Claude** in claude.ai and as **Global instructions** in Cowork, in beta for Claude Enterprise organizations. Call `GET /v1/compliance/organizations/{org_uuid}/user_preferences` with your existing Compliance Access Key and the `read:compliance_user_data` scope. See [List user preferences](https://platform.claude.com/docs/en/manage-claude/compliance-org-data#list-user-preferences).
18
19### October 15, 2026
20
21* [Fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) is out of research preview for Claude Opus 5.5, and you no longer need to request access. Requests to Claude Opus 5.5 no longer require the `fast-mode-2026-02-01` beta header, and requests that still send it keep working. Claude Opus 5.5 now has a default fast mode rate limit in every usage tier. Fast mode for Claude Opus 5 and Claude Opus 4.8 remains in research preview. Keep sending the beta header with fast mode requests to those models.
22
23### October 15, 2026
24
18* [Fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) is out of research preview for Claude Opus 5.5, and you no longer need to request access. Requests to Claude Opus 5.5 no longer require the `fast-mode-2026-02-01` beta header, and requests that still send it keep working. Claude Opus 5.5 now has a default fast mode rate limit in the Start, Build, and Scale usage tiers. Fast mode for Claude Opus 5 and Claude Opus 4.8 remains in research preview. Keep sending the beta header with fast mode requests to those models.
2519* [Fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) for Claude Opus 5.5 is now available on [Claude in Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) for deployments hosted on Azure.
2620
2721### October 14, 2026
agents-and-tools/mcp-tunnels/deploy-compose Changed · +1 / -1 lines
from line 288
288288 cat data/ca.crt
289289 ```
290290
291 The tunnel's status flips to **Active** once a certificate is registered. See [Add a CA certificate](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/console#add-a-ca-certificate).
291 The **Needs certificate** badge next to the tunnel's name goes away once a certificate is registered. See [Add a CA certificate](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/console#add-a-ca-certificate).
292292 </Step>
293293
294294 <Step title="Write the proxy config">
agents-and-tools/mcp-tunnels/quickstart Changed · +1 / -1 lines
from line 107
107107 </Tab>
108108 </Tabs>
109109
110 Back in the Console, on the tunnel detail page, click **Add certificate** and upload `data/ca.crt` (or paste its contents). The tunnel status flips to **Active**.
110 Back in the Console, on the tunnel detail page, click **Add certificate** and upload `data/ca.crt` (or paste its contents). The **Needs certificate** badge next to the tunnel's name goes away.
111111 </Step>
112112
113113 <Step title="Write the sample MCP server">
api/beta/messages/batches Changed · +2 / -0 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.
from line 3972
39723972
39733973 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
39743974
3975 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
3976
39753977 - `"standard"`
39763978
39773979 - `"fast"`
from line 7324
73227324
73237325 - `name: "memory"`
73247326
7325 Name of the tool.
7326
7327 This is how the tool will be called by the model and in `tool_use` blocks.
7328
7329 - `allowed_callers: optional array of "direct" or "code_execution_2025082
7327 Name of the
api/beta/messages/batches/create Changed · +2 / -0 lines
from line 3970
39703970
39713971 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
39723972
3973 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
3974
39733975 - `"standard"`
39743976
39753977 - `"fast"`
api/beta/messages/count_tokens Changed · +2 / -0 lines
from line 3723
37233723
37243724 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
37253725
3726 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
3727
37263728 - `"standard"`
37273729
37283730 - `"fast"`
api/beta/messages/create Changed · +2 / -0 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.
from line 3950
39503950
39513951 The inference speed mode for this request. `"fast"` enables high output-tokens-per-second inference.
39523952
3953 Some models require the `fast-mode-2026-02-01` beta header for any `speed` value. See [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode?sdk_languages=http) for which models.
3954
39533955 - `"standard"`
39543956
39553957 - `"fast"`
from line 7158
71567158
71577159 This is how the tool will be called by the model and in `tool_use` blocks.
71587160
7159 - `allowed_callers: optional array of "direct" or "code_execution_20250825" or "code_execution_20260120" or "code_execution_20260521"`
7160
7161 - `"direct"`
7162
7163 - `"code_execution_20
7161 - `allow
api/compliance/apps/sessions/local/messages Changed · +1 / -1 lines
from line 276
276276
277277 - `updated_at: string`
278278
279 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
279 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
280280
281281 format: date-time
282282
api/compliance/apps/sessions/local/messages/list Changed · +1 / -1 lines
from line 274
274274
275275 - `updated_at: string`
276276
277 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
277 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
278278
279279 format: date-time
280280
api/compliance/apps/sessions/local/retrieve Changed · +1 / -1 lines
from line 60
6060
6161- `updated_at: string`
6262
63 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page or `created_at.lt` window boundary it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
63 Timestamp of the session's last retained inference call (RFC 3339, UTC). Always at or after `created_at`. When a session's activity spans the child organization's retention boundary, calls older than the boundary are no longer reflected — but because retention removes only the oldest calls, this value (unlike `created_at`) is unaffected until the entire session has aged out. On the list endpoint this value is a lower bound: for a session still active at a page boundary or the upper end of a `created_at` window it can momentarily lag the session's true last activity. Retrieving the session, or its messages, always reflects the exact latest retained call.
6464
6565 format: date-time
6666
api/rate-limits Changed · +1 / -1 lines
from line 236
236236
237237### Fast mode rate limits
238238
239When using [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) with `speed: "fast"` on Claude Opus 5.5, Claude Opus 5, or Opus 4.8, dedicated rate limits apply that are separate from standard Opus rate limits. Claude Opus 5.5 has a default fast mode rate limit in every usage tier, so you don't need to request access to use fast mode with it. Fast mode for Claude Opus 5 and Claude Opus 4.8 is in research preview. To see your fast mode rate limits, go to the [Rate limits](https://platform.claude.com/usage/limits) page in the Claude Console. When fast mode rate limits are exceeded, the API returns a `429` error with a `retry-after` header. Fast mode is not available on Claude Opus 4.7 (requests return an error) or Claude Opus 4.6. See [Fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models).
239When using [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) with `speed: "fast"` on Claude Opus 5.5, Claude Opus 5, or Claude Opus 4.8, dedicated rate limits apply that are separate from standard Opus rate limits. Claude Opus 5.5 has a default fast mode rate limit in the Start, Build, and Scale tiers, so you don't need to request access to use fast mode with it. Organizations in the Evaluation tier can use fast mode with Claude Opus 5.5 once they move to the Start, Build, or Scale tier. Fast mode for Claude Opus 5 and Claude Opus 4.8 is in research preview. To see your fast mode rate limits, go to the [Rate limits](https://platform.claude.com/usage/limits) page in the Claude Console. When fast mode rate limits are exceeded, the API returns a `429` error with a `retry-after` header.
240240
241241The response includes `anthropic-fast-*` headers that indicate your fast mode rate limit status. See [Fast mode rate limits](https://platform.claude.com/docs/en/build-with-claude/fast-mode#rate-limits) for details on these headers.
242242
build-with-claude/preserved-thinking Changed · +3 / -1 lines
from line 443
443443<Tip>
444444 **Automate this check with the Claude API skill.** In the latest version of [Claude Code](https://code.claude.com/docs/en/overview), open the repository that builds your requests and run the bundled [Claude API skill](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/claude-api-skill#checking-an-integration-for-preserved-thinking):
445445
446 ```text wrap
446 Let Claude Code check your code for you
447
448 ```text
447449 /claude-api preserved-thinking-migration
448450 ```
449451