HomeDocs changes
Documentation changes
What the docs quietly changed.
Anthropic's published documentation, fetched on a schedule and diffed against the last copy. Every edit dated, including the ones that never made a release note.
Last poll read all 4 cleanly: Claude Code CLI, Claude Developer Platform, Claude Documentation, Model Context Protocol. A date here is when this site noticed, not when Anthropic edited.
Claude Developer Platform
821pages tracked
767 changed this week · last read 24 Oct 2026
Claude Code CLI
228pages tracked
146 changed this week · last read 24 Oct 2026
Claude Documentation
280pages tracked
69 changed this week · last read 24 Oct 2026
Model Context Protocol
357pages tracked
171 changed this week · last read 23 Oct 2026
100 on this page
ReadSourcePageWhatLines
9 Oct 2026
658 pages
+60364 −28978 lines
claude-docs
claude-tag/users/troubleshooting
±changed
+6−6
claude-docschanged
claude-tag/users/troubleshooting
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+6 −6 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/users/models
±changed
+3−3
claude-docschanged
claude-tag/users/models
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+3 −3 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/users/good-habits
±changed
+9−5
claude-docschanged
claude-tag/users/good-habits
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+9 −5 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/users/commands
±changed
+3−3
claude-docschanged
claude-tag/users/commands
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+3 −3 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/admins/healthcare
±changed
+18−20
claude-docschanged
claude-tag/admins/healthcare
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+18 −20 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/admins/for-slack-admins
±changed
+3−3
claude-docschanged
claude-tag/admins/for-slack-admins
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+3 −3 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/admins/customize
±changed
+15−17
claude-docschanged
claude-tag/admins/customize
export const BetaNote = () => <Info>Claude Tag is in public beta. Features and behavior described here may change before general availability.</Info>;
<BetaNote />
> [!NOTE]
>
+15 −17 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 20 Oct 2026 18:07 UTC
claude-docs
claude-tag/users/when-claude-responds
±changed
+1−1
claude-docschanged
claude-tag/users/when-claude-responds
* **Slack Connect channels.** Claude doesn't answer in [channels shared with another company](/docs/claude-tag/admins/restrict-access#slack-connect-channels). A mention there gets
* **Slack Connect channels.** By default, Claude is off in [channels shared with another company](/docs/claude-tag/admins/restrict-access#slack-connect-channels), and a mention the
+1 −1 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/users/troubleshooting
±changed
+20−14
claude-docschanged
claude-tag/users/troubleshooting
### Claude never responds in a channel shared with another company
**What you see**
A mention in a Slack Connect channel, one shared with another company, gets a notice that Claude isn't turned on for Slack Connect channels, and no answer.
+20 −14 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/users/models
±changed
+0−2
claude-docschanged
claude-tag/users/models
Fast mode isn't available in a [channel shared with another company](/docs/claude-tag/users/troubleshooting#claude-never-responds-in-a-channel-shared-with-another-company).
+0 −2 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/users/good-habits
±changed
+1−1
claude-docschanged
claude-tag/users/good-habits
Claude can't post into a channel in a different workspace, and it doesn't answer in a [Slack Connect channel](/docs/claude-tag/admins/restrict-access#slack-connect-channels), one s
Claude can't post into a channel in a different workspace. In a [Slack Connect channel](/docs/claude-tag/admins/restrict-access#slack-connect-channels), one shared with another com
+1 −1 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/users/commands
±changed
+2−2
claude-docschanged
claude-tag/users/commands
In a [Slack Connect channel](/docs/claude-tag/admins/restrict-access#slack-connect-channels), one shared with another company, only `!help`, `!mute`, `!unmute`, `!restart`, and `!f
In a [channel shared with another company](/docs/claude-tag/users/troubleshooting#claude-never-responds-in-a-channel-shared-with-another-company), Claude doesn't answer `!fast`. Us
+2 −2 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/concepts/security-and-data
±changed
+1−1
claude-docschanged
claude-tag/concepts/security-and-data
Claude doesn't work in a channel shared with another company through [Slack Connect](/docs/claude-tag/admins/restrict-access#slack-connect-channels), so no connector's credentials
Connector credentials reach a channel shared with another company through [Slack Connect](/docs/claude-tag/admins/restrict-access#slack-connect-channels) only through a bundle an O
+1 −1 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/admins/restrict-access
±changed
+42−9
claude-docschanged
claude-tag/admins/restrict-access
* **Stop Claude in one workspace or channel:** open the workspace's or channel's page on the **Channels** tab and turn off the **Enable Claude Tag in this workspace** or **Enabl
* **Stop Claude in one workspace or channel:** open the workspace's or channel's page on the **Channels** tab and turn off the **Enable Claude Tag in this workspace** or **Enabl
If someone invites the app into another channel afterward, Claude stays silent there. Mentioning `@Claude` in that channel gets a notice that Claude is disabled in the channel, not
If someone invites the app into another channel afterward, Claude stays silent there. Mentioning `@Claude` in that channel gets a notice that Claude is disabled in the channel, not
* **Shared channels.** A [channel shared across workspaces in your Enterprise Grid](#channels-shared-across-workspaces-in-your-enterprise-grid) takes its settings from the **Slack*
+42 −9 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/admins/healthcare
±changed
+1−1
claude-docschanged
claude-tag/admins/healthcare
Claude never searches private channels, and it doesn't reply in [Slack Connect channels](/docs/claude-tag/admins/restrict-access#slack-connect-channels), the channels your workspac
Claude never searches private channels. In a [Slack Connect channel](/docs/claude-tag/admins/restrict-access#slack-connect-channels), one your workspace shares with another company
+1 −1 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/admins/for-slack-admins
±changed
+1−1
claude-docschanged
claude-tag/admins/for-slack-admins
In a Slack Connect channel (shared with another company), Claude doesn't answer, and a mention there gets a notice saying so. See [Slack Connect channels](/docs/claude-tag/admins/r
In a Slack Connect channel (shared with another company), Claude answers only where an Owner or a Claude Tag admin of your Claude organization has [turned it on](/docs/claude-tag/a
+1 −1 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-docs
claude-tag/admins/customize
±changed
+0−1
claude-docschanged
claude-tag/admins/customize
* **Where**: sessions stay at standard speed in a channel shared with another company
+0 −1 lines
Upstream edited 9 Oct 2026 19:07 UTC Recorded here 9 Oct 2026 19:37 UTC
claude-code
claude-tag
±changed
+2−2
claude-codechanged
claude-tag
No line of the body moved in this read.
+2 −2 lines
Recorded here 9 Oct 2026 19:07 UTC
claude-code
hooks-guide
±changed
+17−14
claude-codechanged
hooks-guide
When the hook succeeds, Claude Code shows nothing in the conversation. To confirm the hook ran, check that the edited file is reformatted, or see [Debug techniques](#debug-techniqu
When the hook succeeds, Claude Code shows nothing in the conversation. To confirm the hook ran, check that the edited file is reformatted, or see [Check what a hook did](#check-wha
The endpoint should return a JSON response body using the same [output format](/docs/en/hooks#json-output) as command hooks. To block a tool call, return a 2xx response with the ap
Your endpoint responds with a JSON body in the same [output format](/docs/en/hooks#json-output) as command hooks, and Claude Code also checks the response status:
+17 −14 lines
Upstream edited 9 Oct 2026 18:44 UTC Recorded here 9 Oct 2026 19:07 UTC
claude-code
hooks
±changed
+111−35
claude-codechanged
hooks
| `onFailure` | no | What happens to the action when the hook fails: `"continue"`, the default, or `"block"`. See [Block the action when a hook fails](#block-the-action-when-a-hook
| `onFailure` | no | What happens to the action when the hook fails: `"continue"`, the default, or `"block"`. See [Block the action when a hook fails](#block-the-action-when-a-hook
The exit code from your hook command tells Claude Code whether the action should proceed, be blocked, or be ignored. The exit code doesn't act alone. Claude Code reads [JSON output
Two tables own the per-event exceptions: [Exit code 2 behavior per event](#exit-code-2-behavior-per-event) says what exit codes do for each event, and [Decision control](#decision-
+111 −35 lines
Upstream edited 9 Oct 2026 18:44 UTC Recorded here 9 Oct 2026 19:07 UTC
claude-code
scheduled-tasks
±changed
+19−6
claude-codechanged
scheduled-tasks
To avoid every session hitting the API at the same wall-clock moment, the scheduler adds a deterministic offset to fire times:
* Recurring tasks fire up to 30 minutes after the scheduled time (or up to half the interval, for tasks that run more often than hourly). An hourly job scheduled for `:00` may fire
* One-shot tasks scheduled for the top or bottom of the hour fire up to 90 seconds early.
+19 −6 lines
Upstream edited 9 Oct 2026 18:40 UTC Recorded here 9 Oct 2026 19:07 UTC
claude-code
claude-tag
±changed
+4−4
claude-codechanged
claude-tag
No line of the body moved in this read.
+4 −4 lines
Recorded here 9 Oct 2026 18:07 UTC
claude-code
interactive-mode
±changed
+6−5
claude-codechanged
interactive-mode
| `f{char}` | Jump to next occurrence of character |
| `F{char}` | Jump to previous occurrence of character |
| `t{char}` | Jump to just before next occurrence of character |
| `T{char}` | Jump to just after previous occurrence of character |
| `f{char}` | Jump to next occurrence of character on the current line |
+6 −5 lines
Upstream edited 9 Oct 2026 17:38 UTC Recorded here 9 Oct 2026 18:07 UTC
claude-code
claude-tag
±changed
+2−2
claude-codechanged
claude-tag
No line of the body moved in this read.
+2 −2 lines
Recorded here 9 Oct 2026 17:19 UTC
claude-code
plugins/manifest-reference
±changed
+6−6
claude-codechanged
plugins/manifest-reference
<Note>
These cases are covered on other pages:
* **Learning to build a plugin**: start with [Create a plugin](/docs/en/plugins/create)
* **What each component does at runtime**: see [Plugin components](/docs/en/plugins/components)
+6 −6 lines
Upstream edited 9 Oct 2026 17:12 UTC Recorded here 20 Oct 2026 18:37 UTC
claude-code
plugins/manifest-reference
±changed
+1−1
claude-codechanged
plugins/manifest-reference
The error reads `Plugin name "<name>" is reserved: it passes as one of Anthropic's own`, and the warning reads `Plugin name "<name>" reads as one of Anthropic's own`. `claude plugi
The error reads `Plugin name "<name>" is reserved: it passes as one of Anthropic's own`, and the warning reads `Plugin name "<name>" reads as one of Anthropic's own`. `claude plugi
+1 −1 lines
Upstream edited 9 Oct 2026 17:12 UTC Recorded here 9 Oct 2026 17:19 UTC
claude-code
agent-view
±changed
+2−2
claude-codechanged
agent-view
`claude attach` and `claude logs` can take part of a running session's name in place of the ID, as in `claude logs "auth refactor"`. Passing a name requires Claude Code v2.1.290 or
`claude attach` and `claude logs` can take part of a session's name in place of the ID, as in `claude logs "auth refactor"`. Passing a name requires Claude Code v2.1.290 or later.
| v2.1.290 | [`claude attach` and `claude logs`](#manage-sessions-from-the-shell) can take part of a running session's name in place of the ID. |
| v2.1.290 | [`claude attach` and `claude logs`](#manage-sessions-from-the-shell) can take part of a session's name in place of the ID. |
+2 −2 lines
Upstream edited 9 Oct 2026 17:09 UTC Recorded here 9 Oct 2026 17:19 UTC
claude-code
claude-tag
±changed
+1−1
claude-codechanged
claude-tag
No line of the body moved in this read.
+1 −1 lines
Recorded here 9 Oct 2026 17:07 UTC
claude-code
plugins/troubleshooting
±changed
+18−0
claude-codechanged
plugins/troubleshooting
<h3 id="a-plugin-stays-installed-after-plugin-uninstall-on-windows">
A plugin stays installed after `plugin uninstall` on Windows
</h3>
+18 −0 lines
Upstream edited 9 Oct 2026 17:03 UTC Recorded here 9 Oct 2026 17:07 UTC
api
release-notes/overview
±changed
+4−0
apichanged
release-notes/overview
### October 9, 2026
* You can now let a Claude Managed Agents agent use [dynamic workflows](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration#dynamic-workflows), in beta with
+4 −0 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/webhooks
±changed
+13−13
apichanged
managed-agents/webhooks
| Event | Trigger
| ---------------------------------- | -----------------------------------------------------------------------------------------------------------------------------------------
| `session.status_run_started` | Agent execution started. This triggers at every session status transition to `running`.
| `session.status_idled` | Agent awaiting input, for example, a tool permission approval or a new user message.
| `session.budget_reached` | The session reached its [budget](https://platform.claude.com/docs/en/managed-agents/budgets) and paused. Fires at most once for each budg
+13 −13 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/tools-web-restrictions
±changed
+17−17
apichanged
managed-agents/tools-web-restrictions
In a [multiagent session](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration), every domain list that applies to a thread is enforced at the same time. An
In a [multiagent session](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration), every domain list that applies to a thread is enforced at the same time. An
* The coordinator's current lists
* The current lists of the agent that the session runs
| Setting | How it combines
+17 −17 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/tools
±changed
+2−0
apichanged
managed-agents/tools
A custom tool's name can't start with `mcp__` or `ant__`, or match the name of a built-in agent tool. A request that breaks this rule returns a 400 error. An agent that already has
+2 −0 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/session-operations
±changed
+9−9
apichanged
managed-agents/session-operations
| Status | Description |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `idle` | Agent is waiting for input, including user messages or tool confirmations. Sessions created without `initial_events` start in `idle`. |
| `running` | Agent is actively executing. |
| `rescheduling` | Transient error occurred, retrying automatically. |
+9 −9 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/session-observability
±changed
+1−1
apichanged
managed-agents/session-observability
Each [session thread](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration)'s own `usage` carries `list_cost` and `active_seconds` too. Per-thread figures ar
Each [session thread](https://platform.claude.com/docs/en/managed-agents/session-threads)'s own `usage` carries `list_cost` and `active_seconds` too. Per-thread figures are rounded
+1 −1 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/self-hosted-sandboxes-workers
±changed
+1−1
apichanged
managed-agents/self-hosted-sandboxes-workers
ARG ANT_VERSION=1.39.1
ARG ANT_VERSION=1.40.0
+1 −1 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/self-hosted-sandboxes-custom-tools
±changed
+1−1
apichanged
managed-agents/self-hosted-sandboxes-custom-tools
| `name` | Unique per agent. Letters, digits, underscores, and hyphens, 1–128 characters. Cannot match a built-in agent tool such as `bash` or `read`, or use the
| `name` | Unique per agent. Letters, digits, underscores, and hyphens, 1–128 characters. Cannot match a built-in agent tool such as `bash` or `read`, or use the
+1 −1 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/reference
±changed
+40−28
apichanged
managed-agents/reference
| `user.interrupt` | Stop the agent mid-execution.
| `user.interrupt` | Stop the agent mid-execution. It ends no [workflow run](https://platform.claude.com/docs/en/managed-agents/workflow-runs#interrupt-a-session-with-
| Type | Description
| -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------
| `agent.message` | Agent response content blocks.
+40 −28 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/quickstart
±changed
+2−2
apichanged
managed-agents/quickstart
VERSION=1.39.1
VERSION=1.40.0
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
+2 −2 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/permission-policies
±changed
+2−2
apichanged
managed-agents/permission-policies
What you post in `user.message` events counts as your intent, and it can lead the server to allow a call it would otherwise deny. The server does not read intent from a tool result
What you post in `user.message` events counts as your intent, and it can lead the server to allow a call it would otherwise deny. The server does not read intent from a tool result
2. The session pauses with a `session.status_idle` event whose `stop_reason.type` is `requires_action`. The blocking event IDs are in the `stop_reason.event_ids` array. The session
2. The session pauses with a `session.status_idle` event whose `stop_reason.type` is `requires_action`. The blocking event IDs are in the `stop_reason.event_ids` array. The session
+2 −2 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/multiagent-orchestration
±changed
+759−1004
apichanged
managed-agents/multiagent-orchestration
## How it works
All agents share the same sandbox, filesystem, and [vault credentials](https://platform.claude.com/docs/en/managed-agents/vaults), but each agent runs in its own **session thread**
Threads are persistent: the coordinator can send a follow-up to an agent it called earlier, and that agent retains everything from its previous turns.
+759 −1004 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/events-and-streaming
±changed
+7−3
apichanged
managed-agents/events-and-streaming
No event resumes a session paused at its budget. The paused work resumes automatically when you change the budget to a value above the consumed list cost, or remove it. See [When a
No event you send resumes a session paused at its budget. The paused work resumes automatically when you change the budget to a value above the consumed list cost, or remove it. Wh
In a multiagent session, a subagent's blocking events are cross-posted to the primary thread. See [Tool permissions and custom tools](https://platform.claude.com/docs/en/managed-ag
In a multiagent session, a subagent's blocking events are cross-posted to the primary thread. See [Tool permissions and custom tools](https://platform.claude.com/docs/en/managed-ag
If the agent has [workflow runs](https://platform.claude.com/docs/en/managed-agents/workflow-runs) open, a custom tool call can arrive while the session stays `running`. The `requi
+7 −3 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/budgets
±changed
+5−3
apichanged
managed-agents/budgets
If [workflow runs](https://platform.claude.com/docs/en/managed-agents/workflow-runs#budgets-and-limits) are open, each one that isn't already idle also gets a `workflow_run.status_
A `user.interrupt` sent while the session is paused at its budget (all threads paused at the cap) is accepted and ignored: it does not appear in the event list and changes nothing.
A `user.interrupt` sent while the session is paused at its budget (all threads paused at the cap) is accepted and ignored: it does not appear in the event list and changes nothing.
Change or remove the budget with a session update. An accepted update resumes the session's paused work automatically; no further client action is needed.
+5 −3 lines
Recorded here 9 Oct 2026 16:37 UTC
api
managed-agents/agent-setup
±changed
+4−4
apichanged
managed-agents/agent-setup
| `multiagent` | A coordinator declaration listing the agents this agent can delegate to. See [Multiagent orchestration](https://platform.claude.com/docs/en/managed-agents/multiag
| `multiagent` | Lets the agent hand work to other agents, as subagents or in [dynamic workflows](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration#dynam
Setting `inference_geo` on a model that doesn't support geographic inference pinning returns a 400 error; see [Model availability](https://platform.claude.com/docs/en/manage-claude
Setting `inference_geo` on a model that doesn't support geographic inference pinning returns a 400 error; see [Model availability](https://platform.claude.com/docs/en/manage-claude
* **`multiagent`** is replaced as a whole, including its `agents` roster. Pass `null` to clear it.
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
get-started
±changed
+2−2
apichanged
get-started
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
<version>2.70.0</version>
<version>2.71.0</version>
+2 −2 lines
Recorded here 9 Oct 2026 16:37 UTC
api
cli-sdks-libraries/sdks/java
±changed
+2−2
apichanged
cli-sdks-libraries/sdks/java
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
<version>2.70.0</version>
<version>2.71.0</version>
+2 −2 lines
Recorded here 9 Oct 2026 16:37 UTC
api
cli-sdks-libraries/cli/sessions-connect
±changed
+1−1
apichanged
cli-sdks-libraries/cli/sessions-connect
The terminal view shows the conversation live: messages and tool calls, with each call's duration and outcome. A status bar shows whether the session is running, idle, or waiting f
The terminal view shows the conversation live: messages and tool calls, with each call's duration and outcome. A status bar shows whether the session is running, idle, or waiting f
+1 −1 lines
Recorded here 9 Oct 2026 16:37 UTC
api
cli-sdks-libraries/cli/quickstart
±changed
+1−1
apichanged
cli-sdks-libraries/cli/quickstart
VERSION=1.39.1
VERSION=1.40.0
+1 −1 lines
Recorded here 9 Oct 2026 16:37 UTC
api
build-with-claude/claude-platform-on-aws
±changed
+4−4
apichanged
build-with-claude/claude-platform-on-aws
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java-aws:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
implementation("com.anthropic:anthropic-java-aws:2.71.0")
<version>2.70.0</version>
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
build-with-claude/claude-on-vertex-ai
±changed
+4−4
apichanged
build-with-claude/claude-on-vertex-ai
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java-vertex:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
implementation("com.anthropic:anthropic-java-vertex:2.71.0")
<version>2.70.0</version>
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
build-with-claude/claude-on-amazon-bedrock-legacy
±changed
+4−4
apichanged
build-with-claude/claude-on-amazon-bedrock-legacy
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java-bedrock:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
implementation("com.anthropic:anthropic-java-bedrock:2.71.0")
<version>2.70.0</version>
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
build-with-claude/claude-in-microsoft-foundry
±changed
+4−4
apichanged
build-with-claude/claude-in-microsoft-foundry
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java-foundry:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
implementation("com.anthropic:anthropic-java-foundry:2.71.0")
<version>2.70.0</version>
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
build-with-claude/claude-in-amazon-bedrock
±changed
+4−4
apichanged
build-with-claude/claude-in-amazon-bedrock
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java-bedrock:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
implementation("com.anthropic:anthropic-java-bedrock:2.71.0")
<version>2.70.0</version>
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
agents-and-tools/tool-use/advisor-tool
±changed
+1−1
apichanged
agents-and-tools/tool-use/advisor-tool
[Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) sessions support an advisor as well, configured as part of the agent rather than as a tool defi
[Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) sessions support an advisor as well, configured as part of the agent rather than as a tool defi
+1 −1 lines
Recorded here 9 Oct 2026 16:37 UTC
api
agents-and-tools/mcp-connector
±changed
+4−4
apichanged
agents-and-tools/mcp-connector
implementation("com.anthropic:anthropic-java:2.70.0")
implementation("com.anthropic:anthropic-java-mcp:2.70.0")
implementation("com.anthropic:anthropic-java:2.71.0")
implementation("com.anthropic:anthropic-java-mcp:2.71.0")
<version>2.70.0</version>
+4 −4 lines
Recorded here 9 Oct 2026 16:37 UTC
api
about-claude/models/optimizing-for-cost-and-intelligence
±changed
+20−20
apichanged
about-claude/models/optimizing-for-cost-and-intelligence
On Claude Managed Agents, you can append the sentence to the task in the `user.message` event that starts the session, and the clock to each `user.message` and [`user.custom_tool_r
On Claude Managed Agents, you can append the sentence to the task in the `user.message` event that starts the session, and the clock to each `user.message` and [`user.custom_tool_r
To use it, add the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) to your request. This beta feature runs the whole strategy server-side
To use it, add the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) to your request. This beta feature runs the whole strategy server-side
To build one, use [multiagent orchestration](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration) in Claude Managed Agents: configure a coordinator agent (t
+20 −20 lines
Recorded here 9 Oct 2026 16:37 UTC
claude-docs
third-party/claude-desktop/admin-console
±changed
+51−1
claude-docschanged
third-party/claude-desktop/admin-console
| **Capabilities** | Whether Chat, Cowork, and Code are each available, the folders and network hosts the app may use, permission modes and built-in tool policy, the [built-in brow
| **Capabilities** | Whether Chat, Cowork, and Code are each available, [Unified Claude](#turn-on-unified-claude), the folders and network hosts the app may use, permission modes a
A group's policy can also turn on Unified Claude for that group, as described under [Turn on Unified Claude](#turn-on-unified-claude).
+51 −1 lines
Upstream edited 9 Oct 2026 15:53 UTC Recorded here 9 Oct 2026 16:07 UTC
api
test-and-evaluate/strengthen-guardrails/reduce-latency
±changed
+1−1
apichanged
test-and-evaluate/strengthen-guardrails/reduce-latency
* **Experiment with temperature:** The `temperature` [parameter](https://platform.claude.com/docs/en/api/messages/create) controls the randomness of the output. Lower values (for e
* **Experiment with temperature:** The `temperature` [parameter](https://platform.claude.com/docs/en/api/messages/create) controls the randomness of the output. Lower values (for e
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
test-and-evaluate/strengthen-guardrails/handle-streaming-refusals
±changed
+3−3
apichanged
test-and-evaluate/strengthen-guardrails/handle-streaming-refusals
* **Fall back to another model:** Configure [server-side fallback or the SDK middleware](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback) so refused req
* **Redeem fallback credit on manual retries:** If you build the retry yourself, pass the refusal's [fallback credit](https://platform.claude.com/docs/en/build-with-claude/fallback
* **Fall back to another model:** Configure [server-side fallback or the SDK middleware](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback) so refused req
* **Redeem fallback credit on manual retries:** If you build the retry yourself, pass the refusal's [fallback credit](https://platform.claude.com/docs/en/build-with-claude/fallback
* **Retry on a different model.** Re-sending a refused request to the same model usually results in another refusal. Instead of only resetting context, retry on a fallback model wi
+3 −3 lines
Recorded here 9 Oct 2026 15:37 UTC
api
resources/overview
±changed
+4−0
apichanged
resources/overview
</Card>
<Card title="Claude Sonnet 5.5 System Card" icon="file" href="https://www.anthropic.com/document/claude-sonnet-5-5-system-card">
Detailed documentation of Claude Sonnet 5.5.
+4 −0 lines
Recorded here 9 Oct 2026 15:37 UTC
api
release-notes/overview
±changed
+1−1
apichanged
release-notes/overview
* Code written for Claude Haiku 4.5 can break on Claude Haiku 5.5. Manual extended thinking (`budget_tokens`) returns a 400 error, and adaptive thinking is on by default, so a resp
* Code written for Claude Haiku 4.5 can break on Claude Haiku 5.5. Manual extended thinking (`budget_tokens`), the sampling parameters `temperature`, `top_p`, and `top_k`, and assi
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/sonnet-5-5/whats-new-sonnet-5-5
±changed
+4−4
apichanged
models/sonnet-5-5/whats-new-sonnet-5-5
* [The advisor tool rejects Claude Opus 4.8, Claude Opus 4.7, and Claude Sonnet 5 as advisors](https://platform.claude.com/docs/en/models/sonnet-5-5/whats-new-sonnet-5-5#advisor-to
* [The advisor tool rejects Claude Opus 4.8, Claude Opus 4.7, Claude Sonnet 5, and Claude Haiku 5.5 as advisors](https://platform.claude.com/docs/en/models/sonnet-5-5/whats-new-son
Every thinking block records which model produced it. Each model reads its own blocks and only some other models' blocks. Claude Sonnet 5.5 reads thinking blocks from Claude Sonnet
Every thinking block records which model produced it. Each model reads its own blocks and only some other models' blocks. Claude Sonnet 5.5 reads thinking blocks from Claude Sonnet
With the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) (beta), a Claude Sonnet 5.5 executor needs Claude Mythos 5.1, Claude Fable 5.1,
+4 −4 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/sonnet-5-5/overview
±changed
+1−1
apichanged
models/sonnet-5-5/overview
* [The advisor tool rejects Claude Opus 4.8, Claude Opus 4.7, and Claude Sonnet 5 as advisors](https://platform.claude.com/docs/en/models/sonnet-5-5/whats-new-sonnet-5-5#advisor-to
* [The advisor tool rejects Claude Opus 4.8, Claude Opus 4.7, Claude Sonnet 5, and Claude Haiku 5.5 as advisors](https://platform.claude.com/docs/en/models/sonnet-5-5/whats-new-son
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/sonnet-5-5/migration-guide
±changed
+2−2
apichanged
models/sonnet-5-5/migration-guide
Claude Sonnet 5.5 reads thinking blocks from Claude Sonnet 5, Claude Opus 4.8, Claude Haiku 4.5, and earlier models. It doesn't read blocks from Claude Opus 5, Claude Opus 5.5, or
Claude Sonnet 5.5 reads thinking blocks from Claude Sonnet 5, Claude Opus 4.8, Claude Haiku 4.5, and earlier models, and, on the Claude API and Google Cloud, from Claude Haiku 5.5.
With the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool), a Claude Sonnet 5.5 executor needs one of these advisors: Claude Opus 5, Claude
With the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool), a Claude Sonnet 5.5 executor needs one of these advisors: Claude Opus 5, Claude
+2 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/overview
±changed
+1−1
apichanged
models/overview
* **Retirement:** Anthropic’s commitment for Anthropic-operated platforms (Claude API, Claude Platform on AWS, Microsoft Foundry). Amazon Bedrock and Google Cloud set their own dat
* **Retirement:** Anthropic’s commitment for the platforms that Anthropic operates. Amazon Bedrock and Google Cloud set their own dates.
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/opus-5-5/whats-new-opus-5-5
±changed
+1−1
apichanged
models/opus-5-5/whats-new-opus-5-5
Every thinking block records which model produced it, and each model reads its own blocks and only some other models'. Claude Opus 5.5 reads thinking blocks from Claude Opus 5, fro
Every thinking block records which model produced it, and each model reads its own blocks and only some other models'. Claude Opus 5.5 reads thinking blocks from Claude Opus 5, fro
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/opus-5-5/migration-guide
±changed
+2−2
apichanged
models/opus-5-5/migration-guide
* If a router or fallback can move a conversation from Claude Opus 5.5 to another model, expect that model to run without Claude Opus 5.5's thinking blocks (Claude Fable 5.1 and Cl
* If a router or fallback can move a conversation from Claude Opus 5.5 to another model, expect that model to run without Claude Opus 5.5's thinking blocks (Claude Fable 5.1 and Cl
On the Claude API, Claude Fable 5.1 and Claude Mythos 5.1 read Claude Opus 5.5 thinking blocks; no other model does. A router or fallback that moves a conversation from Claude Opus
On the Claude API, Claude Fable 5.1 and Claude Mythos 5.1 read Claude Opus 5.5 thinking blocks; no other model does. A router or fallback that moves a conversation from Claude Opus
+2 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/haiku-5-5/whats-new-haiku-5-5
±changed
+24−16
apichanged
models/haiku-5-5/whats-new-haiku-5-5
| Change | Type | Actio
| --------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- | -----
| [Adaptive thinking and effort](https://platform.claude.com/docs/en/models/haiku-5-5/whats-new-haiku-5-5#adaptive-thinking-and-effort) | New | Optio
| [Larger context window and output](https://platform.claude.com/docs/en/models/haiku-5-5/whats-new-haiku-5-5#larger-context-window-and-output) | New | None.
| [Browser use tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-tool) | New | None.
+24 −16 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/haiku-5-5/overview
±changed
+3−2
apichanged
models/haiku-5-5/overview
* Omit `temperature`, `top_p`, and `top_k`, since a non-default value for any of them returns a 400 error.
* Omit `temperature`, `top_p`, and `top_k`. See [Remove sampling parameters](https://platform.claude.com/docs/en/models/haiku-5-5/migration-guide#remove-sampling-parameters) for th
* The minimum cacheable prompt length is 512 tokens. See [Prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#cache-limitations).
1M tokens. How the window is counted and managed.
How the context window is counted and managed.
+3 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/haiku-5-5/migration-guide
±changed
+15−8
apichanged
models/haiku-5-5/migration-guide
description: Switch to Claude Haiku 5.5 from earlier Haiku models with this migration guide. The guidance to enable Claude Haiku 5.5 includes the new model ID, each breaking change
description: Switch to Claude Haiku 5.5 from earlier Haiku models with this migration guide. The guidance to enable Claude Haiku 5.5 includes the new model ID, settings that return
2. Recount your prompts, and revisit `max_tokens` limits and cost estimates, because the same text counts as more tokens. See [Recount tokens](https://platform.claude.com/docs/en/m
2. Recount your prompts, and revisit `max_tokens` limits and cost estimates, because the same text counts as more tokens and large images count as more visual tokens. See [Recount
7. If you use computer use on the Claude API or Google Cloud, move from `computer_20250124` to the `computer_toolset_20260801` toolset. See [Move computer use to the toolset](https
+15 −8 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/haiku-4-5/overview
±changed
+0−8
apichanged
models/haiku-4-5/overview
<Card title="Choosing a model" icon="scales" href="https://platform.claude.com/docs/en/about-claude/models/choosing-a-model">
When to start efficiency-first with Haiku and when to reach for a larger model.
</Card>
+0 −8 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/fable-5/migration-guide
±changed
+2−2
apichanged
models/fable-5/migration-guide
* If you replay conversation history on an earlier model, strip `thinking` and `redacted_thinking` blocks from prior assistant turns first. Thinking blocks from `claude-fable-5` an
* When you switch models, keep passing thinking blocks back unchanged. Thinking blocks from `claude-fable-5` and `claude-mythos-5` are readable only by those two models and by Clau
* If you replay conversation history on an earlier model, strip `thinking` and `redacted_thinking` blocks from prior assistant turns first. Thinking blocks from `claude-fable-5` an
* When you switch models, keep passing thinking blocks back unchanged. Thinking blocks from `claude-fable-5` and `claude-mythos-5` are readable only by those two models and by Clau
+2 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/fable-5-1/whats-new-fable-5-1
±changed
+5−5
apichanged
models/fable-5-1/whats-new-fable-5-1
If you already call Claude Fable 5, three changes are breaking: [forced tool use returns an error](https://platform.claude.com/docs/en/models/fable-5-1/whats-new-fable-5-1#forced-t
If you already call Claude Fable 5, three changes are breaking: [forced tool use returns an error](https://platform.claude.com/docs/en/models/fable-5-1/whats-new-fable-5-1#forced-t
### Earlier models can't read Claude Fable 5.1 thinking blocks
Every thinking block records which model produced it, and it's preserved in one direction only: Claude Fable 5.1 reads earlier models' thinking blocks, and no earlier model reads C
+5 −5 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/fable-5-1/overview
±changed
+1−1
apichanged
models/fable-5-1/overview
If you already call Claude Fable 5, three changes are breaking: [forced tool use returns an error](https://platform.claude.com/docs/en/models/fable-5-1/whats-new-fable-5-1#forced-t
If you already call Claude Fable 5, three changes are breaking: [forced tool use returns an error](https://platform.claude.com/docs/en/models/fable-5-1/whats-new-fable-5-1#forced-t
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
models/fable-5-1/migration-guide
±changed
+7−7
apichanged
models/fable-5-1/migration-guide
Migration is mostly drop-in. The API surface, limits, per-token pricing, tokenizer, always-on adaptive thinking, refusal handling, and `stop_details` categories all match Claude Fa
Migration is mostly drop-in. The API surface, limits, per-token pricing, tokenizer, always-on adaptive thinking, refusal handling, and `stop_details` categories all match Claude Fa
2. **Thinking blocks are preserved only for the model that produced them, or a newer one:** Every `thinking` block records which model produced it. Claude Fable 5.1 reads its own b
A conversation that ran on Claude Fable 5.1 can land on an older model through a router switch, a client-side retry, or a [classifier refusal fallback](https://platform.claude.c
+7 −7 lines
Recorded here 9 Oct 2026 15:37 UTC
api
managed-agents/sessions
±changed
+9−9
apichanged
managed-agents/sessions
"model": {"id": "claude-sonnet-5"},
"model": {"id": "claude-sonnet-5-5"},
id: claude-sonnet-5
id: claude-sonnet-5-5
"model": {"id": "claude-sonnet-5"},
+9 −9 lines
Recorded here 9 Oct 2026 15:37 UTC
api
managed-agents/multiagent-orchestration
±changed
+35−22
apichanged
managed-agents/multiagent-orchestration
model: claude-haiku-4-5
model: claude-haiku-5-5
model: claude-haiku-4-5
model: claude-haiku-5-5
"model": "claude-haiku-4-5",
+35 −22 lines
Recorded here 9 Oct 2026 15:37 UTC
api
managed-agents/events-and-streaming
±changed
+177−189
apichanged
managed-agents/events-and-streaming
3. For each blocking event ID, send a [`user.custom_tool_result`](https://platform.claude.com/docs/en/managed-agents/events-and-streaming#return-a-custom-tool-result) or a [`user.t
3. For each blocking event ID, send a [`user.custom_tool_result`](https://platform.claude.com/docs/en/managed-agents/events-and-streaming#return-a-custom-tool-result) or a [`user.t
You can send the result as soon as the `agent.custom_tool_use` event arrives, without waiting for `session.status_idle`. The session still emits `session.status_idle` with a `requi
+177 −189 lines
Recorded here 9 Oct 2026 15:37 UTC
api
manage-claude/rate-limits-api
±changed
+278−12
apichanged
manage-claude/rate-limits-api
* **`source` on workspace limits:** On the workspace endpoint, every limit value also carries `source`, which says where `value` comes from: `{"type": "workspace"}` for an override
```json
```text wrap
The `/v1/organizations/workspaces/{workspace_id}/rate_limits` endpoint returns the rate limit overrides configured for a single workspace.
+278 −12 lines
Recorded here 9 Oct 2026 15:37 UTC
api
manage-claude/data-residency
±changed
+1−1
apichanged
manage-claude/data-residency
* **Claude 4.6 and later models:** US-only inference (`inference_geo: "us"`) is priced at 1.1x the standard rate across all token pricing categories (input tokens, output tokens, c
* **Claude 4.6 and later models:** US-only inference (`inference_geo: "us"`) is priced at 1.1x the standard rate across all token pricing categories (input tokens, output tokens, c
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
intro
±changed
+1−1
apichanged
intro
Discover what Claude can do: extended thinking, web search, file handling, structured outputs, and more.
Discover what Claude can do: thinking, web search, file handling, structured outputs, and more.
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
cli-sdks-libraries/sdks/java
±changed
+2−2
apichanged
cli-sdks-libraries/sdks/java
Avoid setting a large `maxTokens` value without using streaming. Some networks may drop idle connections after a certain period of time, which can cause the request to fail or [tim
Avoid setting a large `maxTokens` value without using streaming. Some networks may drop idle connections after a certain period of time, which can cause the request to fail or [tim
* **Agent Platform:** `com.anthropic:anthropic-java-vertex`: Use `VertexBackend.fromEnv()` or `VertexBackend.builder()`.
* **Agent Platform:** `com.anthropic:anthropic-java-vertex`: Use `VertexBackend.builder()` or `VertexBackend.fromEnv()`. `fromEnv()` reads `CLOUD_ML_REGION` and `ANTHROPIC_VERTEX_P
+2 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
cli-sdks-libraries/middleware
±changed
+1−1
apichanged
cli-sdks-libraries/middleware
The SDK ships a refusal-fallback middleware that automatically retries requests Claude Fable 5 declines on a fallback model. See [Detect and retry on a fallback model](https://plat
The SDK ships a refusal-fallback middleware that automatically retries a request Claude declines (`stop_reason: "refusal"`) on a fallback model. [Refusals and fallback](https://pla
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
cli-sdks-libraries/libraries/openai-sdk
±changed
+42−42
apichanged
cli-sdks-libraries/libraries/openai-sdk
* System/developer messages are hoisted and concatenated to the beginning of the conversation, as Anthropic only supports a single initial system message.
* System/developer messages are hoisted and concatenated into the top-level `system` field, wherever they appear in the conversation. To add system instructions partway through a c
Most of the inputs to the OpenAI SDK clearly map directly to Anthropic’s API parameters, but one distinct difference is the handling of system / developer prompts. These two prompt
Most of the inputs to the OpenAI SDK clearly map directly to Anthropic’s API parameters, but one distinct difference is the handling of system / developer prompts. These two prompt
| Field | Support status
+42 −42 lines
Recorded here 9 Oct 2026 15:37 UTC
api
cli-sdks-libraries/libraries/apple-foundation-models
±changed
+1−1
apichanged
cli-sdks-libraries/libraries/apple-foundation-models
Pin a Claude [effort level](https://platform.claude.com/docs/en/build-with-claude/effort) for every request with `fixedEffort:`. It takes precedence over the framework's per-reques
Pin a Claude [effort level](https://platform.claude.com/docs/en/build-with-claude/effort) for every request with `fixedEffort:`. It takes precedence over the framework's per-reques
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
claude_api_primer
±changed
+3−3
apichanged
claude_api_primer
Thinking can sometimes help Claude with very hard tasks. The current mechanism is [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) (`thinking: {"
Thinking can sometimes help Claude with very hard tasks. The current mechanism is [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) (`thinking: {"
* Claude Sonnet 5.5 (`claude-sonnet-5-5`, adaptive thinking only, on by default)
* Claude Sonnet 5.5 (`claude-sonnet-5-5`, adaptive thinking only, on by default; `thinking: {"type": "disabled"}` returns a 400 error, so send `thinking: {"type": "between_tools"}`
1. **Tool choice limitation:** Only supports `tool_choice: {"type": "auto"}` (default) or `tool_choice: {"type": "none"}`.
+3 −3 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/working-with-messages
±changed
+1−1
apichanged
build-with-claude/working-with-messages
On Claude Fable 5.1, [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), Claude Fable 5, [Claude Mythos 5](https://platform.claude.com/docs/en/mode
On Claude Fable 5.1, [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), Claude Fable 5, [Claude Mythos 5](https://platform.claude.com/docs/en/mode
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/thinking-troubleshooting
±changed
+5−1
apichanged
build-with-claude/thinking-troubleshooting
Claude Haiku 5.5 returns the same message for a per-message effort change while `thinking: {type: "disabled"}` is set.
To fix it, keep the history append-only: pass earlier turns back exactly as sent and received, add instructions with a [mid-conversation system message](https://platform.claude.com
To fix it, keep the history append-only: pass earlier turns back exactly as sent and received, add instructions with a [mid-conversation system message](https://platform.claude.com
+5 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/thinking-tool-workflows
±changed
+1−1
apichanged
build-with-claude/thinking-tool-workflows
* [Limit tool choice to `auto` or `none` in manual mode](https://platform.claude.com/docs/en/build-with-claude/thinking#thinking-with-tool-use): `tool_choice` options that force to
* [Limit tool choice to `auto` or `none` in manual mode](https://platform.claude.com/docs/en/build-with-claude/thinking#thinking-with-tool-use): `tool_choice` options that force to
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/thinking-steering-and-cost
±changed
+10−4
apichanged
build-with-claude/thinking-steering-and-cost
<Note>
To get less thinking on Claude Sonnet 5.5 and Claude Haiku 5.5, lower the effort level. Asking Claude Sonnet 5.5 in the system prompt to think less doesn't reliably reduce its th
</Note>
You can also steer thinking on a per-message basis from the user turn, independently of the system prompt. Appending `"Please think hard before responding."` to a user message enco
+10 −4 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/thinking
±changed
+7−7
apichanged
build-with-claude/thinking
Claude Haiku 5.5 also has thinking on by default and accepts `thinking: {type: "disabled"}` at [effort](https://platform.claude.com/docs/en/build-with-claude/effort) `high` or belo
Claude Haiku 5.5 also has thinking on by default and accepts `thinking: {type: "disabled"}` at [effort](https://platform.claude.com/docs/en/build-with-claude/effort) `high` or belo
* `display` is invalid with `thinking.type: "disabled"` (there is nothing to display).
* `display` is invalid with `thinking.type: "disabled"` (there is nothing to display) and with `thinking.type: "between_tools"` (Claude Sonnet 5.5 only), which accepts no other fie
1. **Tool choice limitation (manual mode):** tool use with manual extended thinking (`thinking: {type: "enabled"}`) only supports `tool_choice: {"type": "auto"}` (the default) or `
+7 −7 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/streaming
±changed
+2−2
apichanged
build-with-claude/streaming
description: Stream Messages API responses incrementally with server-sent events, including text, tool use, and extended thinking deltas.
description: Stream Messages API responses incrementally with server-sent events, including text, tool use, and thinking deltas.
2. **Handle content types:** Be aware that messages can contain multiple content blocks (`text`, `tool_use`, `thinking`). Tool use and extended thinking blocks cannot be partially
2. **Handle content types:** Be aware that messages can contain multiple content blocks (`text`, `tool_use`, `thinking`). Tool use and thinking blocks cannot be partially recovered
+2 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/refusals-and-fallback
±changed
+10−9
apichanged
build-with-claude/refusals-and-fallback
The simplest setup, in beta on the Claude API: set `fallbacks` to `"default"`, and the API retries a declined request on the fallback model Anthropic recommends for its refusal cat
The simplest setup, in beta on the Claude API: set `fallbacks` to `"default"`, and the API retries a declined request on the fallback model Anthropic recommends for its refusal cat
**Fallback:** When you use fallback, the refusal that triggered it is billed, in addition to the fallback request, when it arrived mid-stream or is in one of the billed categories.
**Fallback:** When you use fallback, the refusal that triggered it is billed, in addition to the fallback request, when it arrived mid-stream or is in one of the billed categories.
Server-side fallback and the SDK middleware apply fallback credit for you. You only need the [Fallback credit](https://platform.claude.com/docs/en/build-with-claude/fallback-credit
+10 −9 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5
±changed
+1−1
apichanged
build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5
Changing the top-level `effort` value between requests invalidates the prompt cache. To run individual turns at a different level, use a [per-message effort change](https://platfor
Changing the top-level `effort` value between requests invalidates the prompt cache. To change the level partway through a conversation, use a [per-message effort change](https://p
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/prompt-engineering/prompting-claude-opus-5-5
±changed
+1−1
apichanged
build-with-claude/prompt-engineering/prompting-claude-opus-5-5
Changing the top-level `effort` value between requests invalidates the prompt cache. To run individual turns at a different level, use a [per-message effort change](https://platfor
Changing the top-level `effort` value between requests invalidates the prompt cache. To change the level partway through a conversation, use a [per-message effort change](https://p
+1 −1 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/prompt-engineering/prompting-claude-haiku-5-5
±changed
+7−3
apichanged
build-with-claude/prompt-engineering/prompting-claude-haiku-5-5
For the five breaking API changes when migrating from Claude Haiku 4.5, see the [migration guide](https://platform.claude.com/docs/en/models/haiku-5-5/migration-guide).
For the breaking API changes when migrating from Claude Haiku 4.5, see the [migration guide](https://platform.claude.com/docs/en/models/haiku-5-5/migration-guide).
Several sections on this page suggest text to add to your system prompt. Add it to new conversations only. A request that sends thinking blocks back after `system` changed can re
* `medium` is the default on the Claude API and in Claude Code. Start here for most work, including agentic coding.
+7 −3 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/prompt-engineering/claude-prompting-best-practices
±changed
+2−2
apichanged
build-with-claude/prompt-engineering/claude-prompting-best-practices
If you are not using extended thinking, no changes are required. On Claude Opus 4.6 through Claude Opus 4.8 and Claude Sonnet 4.6, thinking is off when you omit the `thinking` para
If you are not using extended thinking, you don't need to add a `thinking` parameter. On Claude Opus 4.6 through Claude Opus 4.8 and Claude Sonnet 4.6, thinking is off when you omi
4. **Update thinking configuration:** Claude 4.6 models use [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) (`thinking: {type: "adaptive"}`) ins
4. **Update thinking configuration:** Claude 4.6 and later models use [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) (`thinking: {type: "adapti
+2 −2 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/prompt-caching
±changed
+13−13
apichanged
build-with-claude/prompt-caching
These multipliers stack with other pricing modifiers such as the Batch API discount and data residency. See [pricing](https://platform.claude.com/docs/en/about-claude/pricing) fo
These multipliers stack with other pricing modifiers such as the Batch API discount, [long context pricing](https://platform.claude.com/docs/en/about-claude/pricing#long-context-
| What changes | Tools cache | System cache | Messages cache | Impact
| --------------------------------------------------------- | -------------- | -------------- | -------------- | -------------------------------------------------------------------
| **Tool definitions** | ✘ | ✘ | ✘ | Modifying tool definitions (names, descriptions, parameters) invali
+13 −13 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/preserved-thinking
±changed
+8−6
apichanged
build-with-claude/preserved-thinking
Claude Fable 5.1 and Claude Mythos 5.1 read thinking blocks produced by each other and by earlier Claude models. No earlier model reads thinking blocks from Claude Fable 5.1 or Cla
Claude Fable 5.1 and Claude Mythos 5.1 read thinking blocks produced by each other and by earlier Claude models. No other model reads thinking blocks from Claude Fable 5.1 or Claud
Claude Haiku 5.5 reads thinking blocks from Claude Sonnet 5, Claude Opus 4.8, Claude Haiku 4.5, and earlier models, but not from Claude Opus 5, Claude Opus 5.5, Claude Sonnet 5.5,
* **A conversation that moves down to an earlier model loses Claude Fable 5.1's reasoning for that request.** This occurs when a router sends a turn to a cheaper model, after a [cl
+8 −6 lines
Recorded here 9 Oct 2026 15:37 UTC
api
build-with-claude/mid-conversation-system-messages
±changed
+24−3
apichanged
build-with-claude/mid-conversation-system-messages
* **User input that should not interrupt an agentic loop.** A user types a follow-up while Claude is still executing tools for the previous request. Relaying it as a system message
* **User input that should not interrupt an agentic loop.** A user types a follow-up while Claude is still executing tools for the previous request. Relaying it as a system message
A mid-conversation system message must immediately follow a `user` turn (or an `assistant` turn ending in a server tool result), and must either be the last entry in `messages` or
A mid-conversation system message that carries content (`text`, `tool_addition`, or `tool_removal` blocks) must immediately follow a `user` turn (or an `assistant` turn ending in a
In an [agentic loop](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview), the system message goes after the `user` message that delivers the tool results. This
+24 −3 lines
Recorded here 9 Oct 2026 15:37 UTC
Nothing on this page matches that. Press Enter to search every page of the ledger.