One read of Claude Developer Platformapi-20260824T210708Z
95 pages moved out of 612 read.
Pages moved
95
significant first
Pages read
612
in this capture
Captured
21:07 UTC
Corpus hash
91049975d04a
corpus-hash
What this read moved
1-25 of 95, page 1 of 4This capture is too large to show at once. Changes 1-25 of 95 are below, significant first; the rest are on the following screens.
about-claude/models/choosing-a-model Changed · +4 / -4 lines
from line 60
6060
6161**Claude Opus 5** (`claude-opus-5`) is a step-change improvement over Claude Opus 4.8, strong on deep reasoning, agentic and long-horizon tasks, and test-time compute scaling. Claude Opus 5 supports a 1M token context window by default and up to 128k output tokens, and is priced at $5 USD per million input tokens and $25 USD per million output tokens.
6262
63**Claude Fable 5** (`claude-fable-5`) is Anthropic's most capable widely released model, delivering next-generation intelligence for long-running agents. **Claude Mythos 5** (`claude-mythos-5`) is available through [Project Glasswing](https://anthropic.com/glasswing). Both models support a 1M token context window by default, up to 128k output tokens, and always-on [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking). See [Introducing Claude Fable 5 and Claude Mythos 5](https://platform.claude.com/docs/en/about-claude/models/introducing-claude-fable-5-and-claude-mythos-5) for launch details.
63**Claude Fable 5** (`claude-fable-5`) is Anthropic's most capable widely released model, delivering next-generation intelligence for long-running agents. **Claude Mythos 5** (`claude-mythos-5`) is available through [Project Glasswing](https://anthropic.com/glasswing). Both models support a 1M token context window by default, up to 128k output tokens, and always-on [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking). See [Introducing Claude Fable 5 and Claude Mythos 5](https://platform.claude.com/docs/en/models/fable-5/introducing-claude-fable-5-and-claude-mythos-5) for launch details.
6464
6565Claude Fable 5 and Claude Mythos 5 are priced at $10 USD per million input tokens and $50 USD per million output tokens.
6666
from line 98
9898## Next steps
9999
100100<CardGroup cols={2}>
101 <Card title="Model comparison chart" icon="settings" href="https://platform.claude.com/docs/en/about-claude/models/overview">
101 <Card title="Model comparison chart" icon="settings" href="https://platform.claude.com/docs/en/models/overview">
102102 See detailed specifications and pricing for the latest Claude models
103103 </Card>
104104
105 <Card title="What's new in Claude Opus 5" icon="sparkle" href="https://platform.claude.com/docs/en/about-claude/models/whats-new-opus-5">
105 <Card title="What's new in Claude Opus 5" icon="sparkle" href="https://platform.claude.com/docs/en/models/opus-5/whats-new-opus-5">
106106 Explore the latest improvements in Claude Opus 5
107107 </Card>
108108
109 <Card title="What's new in Claude Sonnet 5" icon="sparkle" href="https://platform.claude.com/docs/en/about-claude/models/whats-new-sonnet-5">
109 <Card title="What's new in Claude Sonnet 5" icon="sparkle" href="https://platform.claude.com/docs/en/models/sonnet-5/whats-new-sonnet-5">
110110 The best combination of speed and intelligence
111111 </Card>
112112
about-claude/models/introducing-claude-fable-5-and-claude-mythos-5 Page removed · 133 lines, page removed
## Models ## Refusals, fallback, and billing on Claude Fable 5 ### Refusals ### Fallback ### Billing ## Availability ## Working with Claude Fable 5 and Claude Mythos 5 ### Prompting ## Messages API on Claude Fable 5 and Claude Mythos 5 ### Adaptive thinking is always on ### Raw thinking content is never returned ## Supported features ## Migrating from earlier models ## Next steps
The page is gone upstream. What it last said is kept here.
about-claude/models/migration-guide Changed · +13 / -13 lines
from line 20
2020
2121## Migrating to Claude Mythos 5 and Claude Fable 5
2222
23[Claude Fable 5](https://platform.claude.com/docs/en/about-claude/models/introducing-claude-fable-5-and-claude-mythos-5) is Anthropic's most capable widely released model, available on the Claude API, [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Google Cloud](https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai), and [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry). [Claude Mythos 5](https://anthropic.com/glasswing) shares the same capabilities and is offered only to approved customers in Project Glasswing.
23[Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/introducing-claude-fable-5-and-claude-mythos-5) is Anthropic's most capable widely released model, available on the Claude API, [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Google Cloud](https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai), and [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry). [Claude Mythos 5](https://anthropic.com/glasswing) shares the same capabilities and is offered only to approved customers in Project Glasswing.
2424
2525The baseline settings shared by `claude-fable-5` and `claude-mythos-5`:
2626
from line 38
3838
3939### Migrating to Claude Mythos 5 and Claude Fable 5 from Claude Mythos Preview
4040
41[Claude Mythos 5](https://anthropic.com/glasswing) is the access-gated successor to [Claude Mythos Preview](https://anthropic.com/glasswing), the invitation-only research preview. [Claude Fable 5](https://platform.claude.com/docs/en/about-claude/models/introducing-claude-fable-5-and-claude-mythos-5) offers the same capabilities and does not require access approval. The changes in this section apply equally to both targets.
41[Claude Mythos 5](https://anthropic.com/glasswing) is the access-gated successor to [Claude Mythos Preview](https://anthropic.com/glasswing), the invitation-only research preview. [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/introducing-claude-fable-5-and-claude-mythos-5) offers the same capabilities and does not require access approval. The changes in this section apply equally to both targets.
4242
4343Migration is mostly drop-in. Claude Mythos 5 and Claude Fable 5 use the same [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) and the same [tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview) patterns as Claude Mythos Preview, and token counts are roughly unchanged because all three models use the same tokenizer. The key changes to check are the features that are no longer available (listed in the next section) and thinking output. If you migrate to Claude Fable 5, also plan for safety classifier refusals, which Claude Mythos Preview and Claude Mythos 5 do not have; see [Refusals and fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback).
4444
from line 330
330330
331331### Migrating to Claude Mythos 5 and Claude Fable 5 from Claude Opus 5
332332
333Claude Fable 5 and Claude Mythos 5 use the same [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) and the same [tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview) patterns as Claude Opus 5, with the same [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default and the same [128k max output tokens](https://platform.claude.com/docs/en/about-claude/models/overview). The prefill and sampling-parameter restrictions, and the thinking display behavior, carry over from Claude Opus 5 unchanged. The changes to check are always-on thinking, pricing, Priority Tier, and data retention.
333Claude Fable 5 and Claude Mythos 5 use the same [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) and the same [tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview) patterns as Claude Opus 5, with the same [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default and the same [128k max output tokens](https://platform.claude.com/docs/en/models/overview). The prefill and sampling-parameter restrictions, and the thinking display behavior, carry over from Claude Opus 5 unchanged. The changes to check are always-on thinking, pricing, Priority Tier, and data retention.
334334
335335#### Update your model name
336336
from line 365
365365 If your code is on Claude Opus 4.7 or earlier, first apply the relevant [Migrating to Claude Opus 5](https://platform.claude.com/docs/en/about-claude/models/migration-guide#migrating-to-claude-opus-5) from-section for the API-level changes from your current model, then the remaining delta in this section.
366366</Note>
367367
368Migration is mostly drop-in. Claude Fable 5 and Claude Mythos 5 use the same [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) and the same [tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview) patterns as Claude Opus 4.8, with the same [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default and the same [128k max output tokens](https://platform.claude.com/docs/en/about-claude/models/overview). Token counts are roughly unchanged because the models use the same tokenizer. The key changes to check are always-on [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), thinking output, safety classifier refusals (Claude Fable 5 only), and pricing.
368Migration is mostly drop-in. Claude Fable 5 and Claude Mythos 5 use the same [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) and the same [tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview) patterns as Claude Opus 4.8, with the same [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default and the same [128k max output tokens](https://platform.claude.com/docs/en/models/overview). Token counts are roughly unchanged because the models use the same tokenizer. The key changes to check are always-on [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), thinking output, safety classifier refusals (Claude Fable 5 only), and pricing.
369369
370370#### Update your model name
371371
from line 701
701701
702702Claude Opus 5 is a step-change improvement over Claude Opus 4.8, strong on deep reasoning, agentic and long-horizon tasks, and test-time compute scaling. For behavioral differences and model-specific prompting patterns, see [Prompting Claude Opus 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5).
703703
704Claude Opus 5 is a drop-in upgrade for Claude Opus 4.8 at the same pricing of $5 per million input tokens and $25 per million output tokens; see [Claude pricing](https://platform.claude.com/docs/en/about-claude/pricing). There are two breaking changes for code already running on Claude Opus 4.8, covered under Breaking changes below. Claude Opus 5 supports the same set of features as Claude Opus 4.8, including the [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) (the default, with no beta header), [128k max output tokens](https://platform.claude.com/docs/en/about-claude/models/overview), [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching), [batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing), the [Files API](https://platform.claude.com/docs/en/build-with-claude/files), [PDF support](https://platform.claude.com/docs/en/build-with-claude/pdf-support), [vision](https://platform.claude.com/docs/en/build-with-claude/vision), and server-side and client-side [tools](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview), with two exceptions: [web fetch](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) is not available on Claude Opus 5, and [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers#supported-models) is not supported on Claude Opus 5. See each tool page for model availability.
704Claude Opus 5 is a drop-in upgrade for Claude Opus 4.8 at the same pricing of $5 per million input tokens and $25 per million output tokens; see [Claude pricing](https://platform.claude.com/docs/en/about-claude/pricing). There are two breaking changes for code already running on Claude Opus 4.8, covered under Breaking changes below. Claude Opus 5 supports the same set of features as Claude Opus 4.8, including the [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) (the default, with no beta header), [128k max output tokens](https://platform.claude.com/docs/en/models/overview), [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching), [batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing), the [Files API](https://platform.claude.com/docs/en/build-with-claude/files), [PDF support](https://platform.claude.com/docs/en/build-with-claude/pdf-support), [vision](https://platform.claude.com/docs/en/build-with-claude/vision), and server-side and client-side [tools](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview), with two exceptions: [web fetch](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) is not available on Claude Opus 5, and [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers#supported-models) is not supported on Claude Opus 5. See each tool page for model availability.
705705
706706### Migrating to Claude Opus 5 from Claude Opus 4.8
707707
from line 794
794794
795795### Migrating to Claude Opus 5 from Claude Opus 4.7
796796
797Claude Opus 5 should have strong out-of-the-box performance on existing Claude Opus 4.7 prompts and evals, at the same pricing of $5 per million input tokens and $25 per million output tokens. It supports the same set of features as Claude Opus 4.7, including the [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows), [128k max output tokens](https://platform.claude.com/docs/en/about-claude/models/overview), [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching), [batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing), the [Files API](https://platform.claude.com/docs/en/build-with-claude/files), [PDF support](https://platform.claude.com/docs/en/build-with-claude/pdf-support), [vision](https://platform.claude.com/docs/en/build-with-claude/vision), and server-side and client-side [tools](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview), with two exceptions: [web fetch](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) is not available on Claude Opus 5, and [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers#supported-models) is not supported on Claude Opus 5. It also adds [mid-conversation system messages](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages) and publicly documents [refusal stop details](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#refusal-response). On the Claude API, Claude Opus 5 also supports [computer use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool) as the stable `computer_toolset_20260801` toolset and the [browser use tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-tool) for tasks inside webpages, neither of which Claude Opus 4.7 supports; existing integrations on the earlier `computer_20251124` version continue to work unchanged on both models. To upgrade an existing integration, see [Migrate from `computer_20251124`](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool#migrate-from-computer-20251124).
797Claude Opus 5 should have strong out-of-the-box performance on existing Claude Opus 4.7 prompts and evals, at the same pricing of $5 per million input tokens and $25 per million output tokens. It supports the same set of features as Claude Opus 4.7, including the [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows), [128k max output tokens](https://platform.claude.com/docs/en/models/overview), [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching), [batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing), the [Files API](https://platform.claude.com/docs/en/build-with-claude/files), [PDF support](https://platform.claude.com/docs/en/build-with-claude/pdf-support), [vision](https://platform.claude.com/docs/en/build-with-claude/vision), and server-side and client-side [tools](https://platform.claude.com/docs/en/agents-and-tools/tool-use/overview), with two exceptions: [web fetch](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) is not available on Claude Opus 5, and [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers#supported-models) is not supported on Claude Opus 5. It also adds [mid-conversation system messages](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages) and publicly documents [refusal stop details](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#refusal-response). On the Claude API, Claude Opus 5 also supports [computer use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool) as the stable `computer_toolset_20260801` toolset and the [browser use tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-tool) for tasks inside webpages, neither of which Claude Opus 4.7 supports; existing integrations on the earlier `computer_20251124` version continue to work unchanged on both models. To upgrade an existing integration, see [Migrate from `computer_20251124`](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool#migrate-from-computer-20251124).
798798
799799<Note>
800800 If your code is on Claude Opus 4.6 or earlier, use [Migrating to Claude Opus 5 from Claude Opus 4.6 and earlier Opus models](https://platform.claude.com/docs/en/about-claude/models/migration-guide#migrating-from-claude-opus-46) instead. That section includes breaking changes (sampling parameters rejected, manual extended thinking rejected, new tokenizer) that the upgrade from Claude Opus 4.7 alone does not cover.
from line 906
906906Claude Opus 5 should have strong out-of-the-box performance on existing Claude Opus 4.6 prompts and evals at the same pricing, but there are a handful of behavioral and API changes worth knowing about as you migrate. Most of these changes took effect in Claude Opus 4.7; two more, thinking on by default and an effort cap on disabling thinking, take effect on Claude Opus 5. All of them are covered below, so this section is complete for code coming straight from Claude Opus 4.6. Claude Opus 5 supports the same set of features as Claude Opus 4.6, including:
907907
908908* [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) at standard API pricing with no long-context premium
909* [128k max output tokens](https://platform.claude.com/docs/en/about-claude/models/overview)
909* [128k max output tokens](https://platform.claude.com/docs/en/models/overview)
910910* [Adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking)
911911* [Prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching)
912912* [Batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing)
from line 2012
20122012
20132013### Migrating to Claude Opus 5 from Claude Sonnet 5
20142014
2015Claude Opus 5 and Claude Sonnet 5 share the same API surface: both run with [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) on by default, both default the [effort parameter](https://platform.claude.com/docs/en/build-with-claude/effort) to `high` on the Claude API and Claude Code, both serve a [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default with [128k max output tokens](https://platform.claude.com/docs/en/about-claude/models/overview), and neither supports [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers#supported-models). Manual extended thinking and non-default sampling parameters return a 400 error on both models, as does assistant prefill.
2015Claude Opus 5 and Claude Sonnet 5 share the same API surface: both run with [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) on by default, both default the [effort parameter](https://platform.claude.com/docs/en/build-with-claude/effort) to `high` on the Claude API and Claude Code, both serve a [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default with [128k max output tokens](https://platform.claude.com/docs/en/models/overview), and neither supports [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers#supported-models). Manual extended thinking and non-default sampling parameters return a 400 error on both models, as does assistant prefill.
20162016
20172017#### Update your model name
20182018
from line 2027
20272027
202820282. **Disabling thinking is capped at `high` effort:** On Claude Sonnet 5, `thinking: {type: "disabled"}` is accepted at any effort level. On Claude Opus 5, it is accepted only at an [effort](https://platform.claude.com/docs/en/build-with-claude/effort) level of `high` or below; a request that combines `thinking: {type: "disabled"}` with effort `xhigh` or `max` returns a 400 error, enforced on each request. Audit requests that disable thinking before you migrate.
20292029
20303. **Mid-conversation system messages:** Claude Opus 5 accepts `role: "system"` messages immediately after a user turn in the `messages` array (subject to [placement rules](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages#limitations)). This feature is not available on Claude Sonnet 5; use the top-level `system` field instead. If you maintain code paths that rebuild the full message history to update instructions, you can simplify them and preserve [prompt cache](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) hits on earlier turns.
20303. **Mid-conversation system messages:** Claude Opus 5 accepts `role: "system"` messages immediately after a user turn in the `messages` array (subject to [placement rules](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages#limitations)). This feature is not available on Claude Sonnet 5. If you maintain code paths that rebuild the full message history to update instructions, you can simplify them and preserve [prompt cache](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) hits on earlier turns.
20312031
203220324. **Web fetch is not available:** The [web fetch](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) tool is available on Claude Sonnet 5 but not on Claude Opus 5.
20332033
from line 2651
26512651
26522652### Migrating to Claude Sonnet 5 from Claude Haiku 4.5
26532653
2654Claude Haiku 4.5 and Claude Sonnet 5 differ more at the API level than adjacent models within one class: Claude Haiku 4.5 uses manual [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) (off by default), a 200k token context window, and up to 64k output tokens, while Claude Sonnet 5 runs with [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) on by default, serves a [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default, and supports up to [128k output tokens](https://platform.claude.com/docs/en/about-claude/models/overview).
2654Claude Haiku 4.5 and Claude Sonnet 5 differ more at the API level than adjacent models within one class: Claude Haiku 4.5 uses manual [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) (off by default), a 200k token context window, and up to 64k output tokens, while Claude Sonnet 5 runs with [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) on by default, serves a [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default, and supports up to [128k output tokens](https://platform.claude.com/docs/en/models/overview).
26552655
26562656#### Update your model name
26572657
from line 2691
26912691
26922692Claude Haiku 4.5 is the fastest and most intelligent Haiku model with near-frontier performance, delivering premium model quality for interactive applications and high-volume processing.
26932693
2694For a complete overview of capabilities, see the [models overview](https://platform.claude.com/docs/en/about-claude/models/overview).
2694For a complete overview of capabilities, see the [models overview](https://platform.claude.com/docs/en/models/overview).
26952695
26962696<Note>
26972697 For Claude Haiku 4.5 pricing, see [Claude pricing](https://platform.claude.com/docs/en/about-claude/pricing).
from line 2719
27192719
27202720**Review new rate limits:** Haiku 4.5 has separate rate limits from Haiku 3.5. See [Rate limits](https://platform.claude.com/docs/en/api/rate-limits) documentation for details.
27212721
2722**Explore new capabilities:** See the [models overview](https://platform.claude.com/docs/en/about-claude/models/overview) for details on context awareness, increased output capacity (64k tokens), higher intelligence, and improved speed.
2722**Explore new capabilities:** See the [models overview](https://platform.claude.com/docs/en/models/overview) for details on context awareness, increased output capacity (64k tokens), higher intelligence, and improved speed.
27232723
27242724#### Breaking changes
27252725
from line 2766
27662766## Get help
27672767
27682768* Check the [API documentation](https://platform.claude.com/docs/en/api/overview) for detailed specifications
2769* Review [model capabilities](https://platform.claude.com/docs/en/about-claude/models/overview) for performance comparisons
2769* Review [model capabilities](https://platform.claude.com/docs/en/models/overview) for performance comparisons
27702770* Review [API release notes](https://platform.claude.com/docs/en/release-notes/api) for API updates
27712771* Contact support if you encounter any issues during migration
27722772
about-claude/models/optimizing-for-cost-and-intelligence Changed · +14 / -2 lines
### Upgrade the model
from line 21
2121| ------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
2222| Any workload, any model | Turn on prompt caching and trim unneeded tokens; both are free | [Cache repeated context](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#cache-repeated-context) · [Trim tokens](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#trim-input-and-context-tokens) |
2323| Costs are too high; quality is fine | Sweep effort down on your current model | [Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort) |
24| You are not on the latest model | Upgrade; the current model solves more, usually at lower cost per solved task | [Upgrade the model](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#upgrade-the-model) |
2425| You are choosing or switching models | Compare on cost per completed task, not per token | [Compare models](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task) |
2526| Quality isn't good enough | If you lowered effort, restore it; otherwise try the next tier up at `low` effort | [Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort) · [Compare models](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task) |
2627| Attempts end with `stop_reason: max_tokens` | Raise `max_tokens`; 64,000 covered every turn measured and cost nothing extra per solved task | [Set budgets](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#set-budgets-and-output-caps) |
from line 177
176177
177178## Trade cost against intelligence
178179
179These levers set where a single model sits between cost and intelligence: model choice, effort, re-running failures at a higher setting, and the budgets and caps it works within. Start with an effort sweep on your current model ([Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort)). From lowest to highest cost and capability, the current models are Claude Haiku 4.5, Claude Sonnet 5, Claude Opus 5, and Claude Fable 5 (the frontier model); [Models overview](https://platform.claude.com/docs/en/about-claude/models/overview) has the full lineup and prices.
180These levers set where a single model sits between cost and intelligence: model choice, effort, re-running failures at a higher setting, and the budgets and caps it works within. Start with an effort sweep on your current model ([Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort)). From lowest to highest cost and capability, the current models are Claude Haiku 4.5, Claude Sonnet 5, Claude Opus 5, and Claude Fable 5 (the frontier model); [Models overview](https://platform.claude.com/docs/en/models/overview) has the full lineup and prices.
180181
181182### Compare models on cost per task
182183
from line 199
198199
199200The [multi-model strategies](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#combine-models) exist to spend frontier intelligence on that tail without paying frontier rates for the rest.
200201
202### Upgrade the model
203
204If you are a model or two behind, the cheapest lever is the model string. Anthropic ran recent Claude Opus and Claude Sonnet models through the same harness on the SWE-bench Pro[3](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) subset, each at its shipped defaults, and priced each at list rates:
205
206
207
208Anthropic prices the Opus line identically per token across versions, so any saving comes from efficiency: Opus 4.8 takes fewer turns and rereads less than Opus 4.7, so each solved task costs 13% less, and Opus 5 then solves 11 more points of tasks at the same cost. Sonnet 5's savings come from its lower per-token price, which more than offsets the extra tokens it uses per task compared with Sonnet 4.6.
209
210Compare on cost per solved task, not per token: the same text costs about 30% more tokens on Claude Opus 4.7 and later, so a per-token comparison makes the newer models look more expensive by construction.
211
201212### Tune effort
202213
203214Effort is the most direct way to tune a model to your task. The `effort` parameter governs how much thinking, tool calling, and self-verification the model does, and the default (`high`) suits demanding tasks. Cost scales with all that activity; accuracy scales only with the part your task needs. Below the model's ceiling, the highest effort levels pay for depth the task never uses.
from line 619
608619| Data files through code execution | 92% on a 25-question data task | A gain, 25 of 25 instead of 6 of 25 | Faster | [Trim input and context tokens](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#trim-input-and-context-tokens) |
609620| Batch API | 50% | None | Results within 24 hours | [Batch work that can wait](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#batch-work-that-can-wait) |
610621| Prompt audit against the current model | 14% on both migrations measured | None; a gain on one | Faster (fewer tool rounds) | [Audit prompts against the current model](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#audit-prompts-against-the-current-model) |
622| Upgrade the model | Opus 4.7 to Opus 5: about 12% less per solved task, 11 more points; Sonnet 4.6 to Sonnet 5: 14% less, 5 more points | A gain | Neutral | [Upgrade the model](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#upgrade-the-model) |
611623| Lower effort | Knowledge work: `medium` 15% to 30%, `low` a third to a half; long coding: `medium` about half, `low` about three quarters | 1 to 3 points on knowledge work, 2 to 8 on long coding | Faster | [Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort) |
612624| Re-run failures | About half, at the same pass rate | None | Two runs on the tasks that fail | [Re-run failures at higher effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#re-run-failures-at-higher-effort) |
613625| Task budget | 18% to 47% | 3 to 4 points | Faster | [Set budgets and output caps](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#set-budgets-and-output-caps) |
from line 634
622634
6236351. **WideSearch:** Wong et al., "WideSearch: Benchmarking Agentic Broad Info-Seeking," arXiv:2508.07999, 2025. Broad web-research tasks graded on a many-row table's completeness and accuracy; 200 problems, 3 runs per configuration. The caching and effort charts come from separate runs, so per-problem costs differ slightly. The cost-concentration chart is a separate 20-problem run, 3 runs per problem, costed from per-request billing records.
6246362. **GDPval:** OpenAI, "GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks," 2025. Knowledge-work deliverables graded against task rubrics; a 210-task run of the released gold set, one attempt per task. A Claude model grades, so absolute scores may differ from published results.
6253. **SWE-bench Pro:** Scale AI, "SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?", 2025. A 482-problem subset selected for compatibility with Anthropic's evaluation harness; scores are not comparable to the public leaderboard. Claude Opus 5 at the default effort averages two runs; reduced-effort settings are single runs. Escalation figures come task by task from those runs: `low` first, then the default on its failures, solved 92.5% to 93.6% across run pairings for about $0.70; `medium` first, 93.8% to 94.2% for about $0.95; the default re-run on its own failures, 94.0% for $1.58; everything at the default, 90.9% to 92.5% for $1.39. The Claude Sonnet 5 executor pairings on the advisor chart come from the same August 2026 series on this subset: the Sonnet-plus-Opus pairing was run twice (a run and an exact replication), the low-effort pairing once, and Claude Sonnet 5 alone twice (77.4%, the baseline for both Pro rows); the task-budget figures are one run per budget on the same subset. The Claude Fable 5 figure in [Compare models](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task) is a single July 2026 run, also the task-budget chart's unbudgeted baseline; every budgeted run completed all 482 problems without harness errors.
6373. **SWE-bench Pro:** Scale AI, "SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?", 2025. A 482-problem subset selected for compatibility with Anthropic's evaluation harness; scores are not comparable to the public leaderboard. Claude Opus 5 at the default effort averages two runs; reduced-effort settings are single runs. Escalation figures come task by task from those runs: `low` first, then the default on its failures, solved 92.5% to 93.6% across run pairings for about $0.70; `medium` first, 93.8% to 94.2% for about $0.95; the default re-run on its own failures, 94.0% for $1.58; everything at the default, 90.9% to 92.5% for $1.39. The Claude Sonnet 5 executor pairings on the advisor chart come from the same August 2026 series on this subset: the Sonnet-plus-Opus pairing was run twice (a run and an exact replication), the low-effort pairing once, and Claude Sonnet 5 alone twice (77.4%, the baseline for both Pro rows); the task-budget figures are one run per budget on the same subset. The Claude Fable 5 figure in [Compare models](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task) is a single July 2026 run, also the task-budget chart's unbudgeted baseline; every budgeted run completed all 482 problems without harness errors. The upgrade ladder is one run per model at its shipped defaults (two each for Opus 5 and Sonnet 5), run the same week in one harness and organization.
6266384. **BrowseComp:** Wei et al., "BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents," OpenAI, 2025. Effort figures use a 500-problem cut, one to three runs per setting. The cost-insurance chart uses 10 reliably solved problems from a 26-problem slice, 50 delegated and 70 solo runs ($6.45 compared with $11.99 per run in expectation); delegated figures carry a measurement band of about 20%.
6276395. **Agent-architecture scaling:** Kim et al., "Towards a Science of Scaling Agent Systems," arXiv:2512.08296, 2025. Independent external study, cited only for the direction of the finding on when delegation does not pay, not for any figure.
6286406. **DeepWideSearch:** "DeepWideSearch: Benchmarking Depth and Width in Agentic Information Seeking," arXiv:2510.20168, 2025. The 220 questions span 15 domains, each combining many-row collection with multi-hop retrieval; measured on the benchmark's standing row set, 3 runs per configuration.
about-claude/models/overview Page removed · 183 lines, page removed
## Choosing a model ### Claude Fable 5 and Claude Mythos 5 ### Latest models comparison ## Prompt and output performance ## Migrating to Claude Opus 5 ## Get started with Claude
The page is gone upstream. What it last said is kept here.
about-claude/models/whats-new-opus-5 Page removed · 351 lines, page removed
## New model ## New features ### Mid-conversation tool changes (beta) ### Default fallbacks mode ### Lower prompt cache minimum ### Fast mode ## Behavior changes ### Thinking on by default ### Effort matters more ### Disabling thinking requires effort `high` or below ### Model behavior differences ## Capability improvements ## Pricing ## Availability ## Migration guide ## Next steps
The page is gone upstream. What it last said is kept here.
about-claude/models/whats-new-sonnet-5 Page removed · 216 lines, page removed
## New model ## Behavior changes ### Adaptive thinking on by default ### Sampling parameters not accepted ### Manual extended thinking removed ## New tokenizer ## API constraints inherited from Claude Sonnet 4.6 ### Assistant message prefilling not supported ## Capability improvements ## Cybersecurity safeguards ## Pricing ## Availability ## Migration guide ## Next steps
The page is gone upstream. What it last said is kept here.
api/compliance Changed · +104447 / -30540 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/compliance/activities Changed · +104368 / -30512 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/compliance/activities/list Changed · +52228 / -15260 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/compliance/apps Changed · +73 / -26 lines
from line 15
1515compliance review. Results are sorted chronologically (time ascending)
1616by the `order_by` key, with ties broken by id.
1717
18**Deprecation notice:** Combining `user_ids[]` with any `updated_at.*`
19filter is deprecated and will be rejected with HTTP 400 after
202026-09-22. For incremental polling by update time, omit `user_ids[]`
21and set `order_by=updated_at` with `after_id` cursor pagination —
22this returns the same chats across the whole organization in a single
23request stream. For per-user listing, use `created_at.*` filters (or
24no time filter) with the default `order_by`. `user_ids[]` with
25`order_by=updated_at` is already rejected.
26
1827### Query Parameters
1928
2029- `after_id: optional string`
from line 76
6776
6877 - `gt: optional string`
6978
70 Filter chats updated after this time (RFC 3339 format)
79 Filter chats updated after this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7180
7281 - `gte: optional string`
7382
74 Filter chats updated at or after this time (RFC 3339 format)
83 Filter chats updated at or after this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7584
7685 - `lt: optional string`
7786
78 Filter chats updated before this time (RFC 3339 format)
87 Filter chats updated before this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7988
8089 - `lte: optional string`
8190
82 Filter chats updated at or before this time (RFC 3339 format)
91 Filter chats updated at or before this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
8392
8493- `user_ids: optional array of string`
8594
86 Filter to chats created by specific users (max 10 per request). Omit for an org-wide query. Enumerate IDs via `GET /v1/compliance/organizations/{org_uuid}/users`.
95 Filter to chats created by specific users (max 10 per request). Omit for an org-wide query. Enumerate IDs via `GET /v1/compliance/organizations/{org_uuid}/users`. Deprecated combination: passing `user_ids[]` together with any `updated_at.*` filter is deprecated and will be rejected after 2026-09-22. For `updated_at`-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
8796
8897### Header Parameters
8998
from line 122
113122
114123 - `model: string or null`
115124
116 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
125 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
117126
118127 - `name: string`
119128
from line 186
177186 "created_at": "2025-06-07T08:09:10Z",
178187 "updated_at": "2025-06-07T09:10:11Z",
179188 "organization_id": "org_abc123",
180 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
189 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
181190 "project_id": "claude_proj_xyz789",
182 "model": "claude-opus-4-7",
191 "model": "claude-opus-5",
183192 "user": {
184193 "id": "user_xyz456",
from line 274
265274
266275 - `model: string or null`
267276
268 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
277 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
269278
270279 - `name: string`
271280
from line 632
623632
624633- `model: string or null`
625634
626 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
635 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
627636
628637- `name: string`
629638
from line 682
673682 "created_at": "2025-06-07T08:09:10Z",
674683 "updated_at": "2025-06-07T08:09:11Z",
675684 "organization_id": "org_abc123",
676 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
685 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
677686 "project_id": "claude_proj_xyz789",
678 "model": "claude-opus-4-7",
687 "model": "claude-opus-5",
679688 "user": {
680689 "id": "user_xyz456",
from line 2789
27802789
27812790 Opaque pagination token from a previous response's `next_page` field. Pass this to retrieve the next page of results. Clients should treat this value as an opaque string and not attempt to parse or interpret its contents, as the format may change without notice.
27822791
2792- `updated_at: optional object { gte }`
2793
2794 - `gte: optional string`
2795
2796 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.
2797
27832798### Header Parameters
27842799
27852800- `"x-api-key": optional string`
from line 2801
27862801
27872802### Returns
27882803
2789- `data: array of object { id, created_at, organization_uuid, 4 more }`
2804- `data: array of object { id, created_at, organization_uuid, 5 more }`
27902805
2791 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order.
2806 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order. `updated_at` never participates in the ordering; the `updated_at.gte` query parameter filters on it without changing the order or the pagination cursor.
27922807
27932808 - `id: string`
27942809
from line 2825
28102825
28112826 - `"compliance_local_session"`
28122827
2828 - `updated_at: string`
2829
2830 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.
2831
28132832 - `user: object { id, email_address }`
28142833
28152834 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 2871
28532872 },
28542873 "product_surface": "cowork",
2855 "created_at": "2026-07-09T14:02:11Z"
2874 "created_at": "2026-07-09T14:02:11Z",
2875 "updated_at": "2026-07-09T15:47:33Z"
28562876 }
28572877 ]
28582878}
from line 2919
28992919
29002920 - `"compliance_local_session"`
29012921
2922- `updated_at: string`
2923
2924 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.
2925
29022926- `user: object { id, email_address }`
29032927
29042928 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 2959
29362960 },
29372961 "product_surface": "cowork",
2938 "created_at": "2026-07-09T14:02:11Z"
2962 "created_at": "2026-07-09T14:02:11Z",
2963 "updated_at": "2026-07-09T15:47:33Z"
29392964}
29402965```
29412966
from line 2968
29432968
29442969### Local List Response
29452970
2946- `LocalListResponse object { id, created_at, organization_uuid, 4 more }`
2971- `LocalListResponse object { id, created_at, organization_uuid, 5 more }`
29472972
29482973 A Cowork or Claude Code session that a user ran on their own computer
29492974 while signed in with their organization account.
from line 2993
29682993
29692994 - `"compliance_local_session"`
29702995
2996 - `updated_at: string`
2997
2998 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.
2999
29713000 - `user: object { id, email_address }`
29723001
29733002 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 3015
29863015
29873016### Local Retrieve Response
29883017
2989- `LocalRetrieveResponse object { id, created_at, organization_uuid, 4 more }`
3018- `LocalRetrieveResponse object { id, created_at, organization_uuid, 5 more }`
29903019
29913020 A Cowork or Claude Code session that a user ran on their own computer
29923021 while signed in with their organization account.
from line 3040
30113040
30123041 - `"compliance_local_session"`
30133042
3043 - `updated_at: string`
3044
3045 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.
3046
30143047 - `user: object { id, email_address }`
30153048
30163049 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 3112
30793112
30803113### Returns
30813114
3082- `data: array of object { id, content, created_at, 3 more }`
3115- `data: array of object { id, content, created_at, 4 more }`
30833116
3084 Transcript turns for this page, ordered by `created_at` in the direction selected by the `order` parameter (ascending by default). Turns sharing a `created_at` (all messages of one inference call carry the call's timestamp) are returned in transcript order.
3117 Transcript turns for this page, in call order: oldest call first by default, newest call first with `order=desc`. The messages of one call carry the call's timestamp and follow each other in transcript order; a page boundary can fall between them.
30853118
30863119 - `id: string`
30873120
from line 3204
31713204
31723205 When the message was recorded (RFC 3339, UTC)
31733206
3207 - `model: string or null`
3208
3209 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
3210
31743211 - `provenance: object { reason, type } or object { type } or object { type } or null`
31753212
31763213 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 3218
31813218
31823219 - `reason: string`
31833220
3184 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
3221 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
31853222
31863223 - `type: "content_unavailable"`
31873224
from line 3268
32313268
32323269 Opaque pagination cursor (prefixed `page_`) for the next page. Null when there is no further page. Treat as an opaque string; the format may change without notice.
32333270
3234- `session: object { id, created_at, organization_uuid, 4 more }`
3271- `session: object { id, created_at, organization_uuid, 5 more }`
32353272
32363273 The local session the messages belong to. `user.email_address` is always null on this endpoint; the messages endpoint does not resolve email addresses.
32373274
from line 3292
32553292
32563293 - `"compliance_local_session"`
32573294
3295 - `updated_at: string`
3296
3297 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.
3298
32583299 - `user: object { id, email_address }`
32593300
32603301 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 3334
32933334 }
32943335 ],
32953336 "created_at": "2025-03-12T18:22:41.123456Z",
3337 "model": "claude-opus-5",
32963338 "provenance": {
32973339 "reason": "not_captured",
32983340 "type": "content_unavailable"
from line 3350
33083350 "organization_uuid": "a1b2c3d4-e5f6-4789-a012-3456789abcde",
33093351 "product_surface": "cowork",
33103352 "type": "compliance_local_session",
3353 "updated_at": "2025-03-12T18:22:41.123456Z",
33113354 "user": {
33123355 "id": "user_01WCz1FkmYMm4gnmykNKUu3Q",
from line 3364
33213364
33223365### Message List Response
33233366
3324- `MessageListResponse object { id, content, created_at, 3 more }`
3367- `MessageListResponse object { id, content, created_at, 4 more }`
33253368
33263369 A single user or assistant turn in a local session transcript.
33273370
from line 3456
34133456
34143457 When the message was recorded (RFC 3339, UTC)
34153458
3459 - `model: string or null`
3460
3461 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
3462
34163463 - `provenance: object { reason, type } or object { type } or object { type } or null`
34173464
34183465 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 3470
34233470
34243471 - `reason: string`
34253472
3426 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
3473 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
34273474
34283475 - `type: "content_unavailable"`
34293476
api/compliance/apps/chats Changed · +21 / -12 lines
from line 13
1313compliance review. Results are sorted chronologically (time ascending)
1414by the `order_by` key, with ties broken by id.
1515
16**Deprecation notice:** Combining `user_ids[]` with any `updated_at.*`
17filter is deprecated and will be rejected with HTTP 400 after
182026-09-22. For incremental polling by update time, omit `user_ids[]`
19and set `order_by=updated_at` with `after_id` cursor pagination —
20this returns the same chats across the whole organization in a single
21request stream. For per-user listing, use `created_at.*` filters (or
22no time filter) with the default `order_by`. `user_ids[]` with
23`order_by=updated_at` is already rejected.
24
1625### Query Parameters
1726
1827- `after_id: optional string`
from line 74
6574
6675 - `gt: optional string`
6776
68 Filter chats updated after this time (RFC 3339 format)
77 Filter chats updated after this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
6978
7079 - `gte: optional string`
7180
72 Filter chats updated at or after this time (RFC 3339 format)
81 Filter chats updated at or after this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7382
7483 - `lt: optional string`
7584
76 Filter chats updated before this time (RFC 3339 format)
85 Filter chats updated before this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7786
7887 - `lte: optional string`
7988
80 Filter chats updated at or before this time (RFC 3339 format)
89 Filter chats updated at or before this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
8190
8291- `user_ids: optional array of string`
8392
84 Filter to chats created by specific users (max 10 per request). Omit for an org-wide query. Enumerate IDs via `GET /v1/compliance/organizations/{org_uuid}/users`.
93 Filter to chats created by specific users (max 10 per request). Omit for an org-wide query. Enumerate IDs via `GET /v1/compliance/organizations/{org_uuid}/users`. Deprecated combination: passing `user_ids[]` together with any `updated_at.*` filter is deprecated and will be rejected after 2026-09-22. For `updated_at`-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
8594
8695### Header Parameters
8796
from line 120
111120
112121 - `model: string or null`
113122
114 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
123 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
115124
116125 - `name: string`
117126
from line 184
175184 "created_at": "2025-06-07T08:09:10Z",
176185 "updated_at": "2025-06-07T09:10:11Z",
177186 "organization_id": "org_abc123",
178 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
187 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
179188 "project_id": "claude_proj_xyz789",
180 "model": "claude-opus-4-7",
189 "model": "claude-opus-5",
181190 "user": {
182191 "id": "user_xyz456",
from line 272
263272
264273 - `model: string or null`
265274
266 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
275 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
267276
268277 - `name: string`
269278
from line 630
621630
622631- `model: string or null`
623632
624 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
633 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
625634
626635- `name: string`
627636
from line 680
671680 "created_at": "2025-06-07T08:09:10Z",
672681 "updated_at": "2025-06-07T08:09:11Z",
673682 "organization_id": "org_abc123",
674 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
683 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
675684 "project_id": "claude_proj_xyz789",
676 "model": "claude-opus-4-7",
685 "model": "claude-opus-5",
677686 "user": {
678687 "id": "user_xyz456",
api/compliance/apps/chats/list Changed · +17 / -8 lines
from line 11
1111compliance review. Results are sorted chronologically (time ascending)
1212by the `order_by` key, with ties broken by id.
1313
14**Deprecation notice:** Combining `user_ids[]` with any `updated_at.*`
15filter is deprecated and will be rejected with HTTP 400 after
162026-09-22. For incremental polling by update time, omit `user_ids[]`
17and set `order_by=updated_at` with `after_id` cursor pagination —
18this returns the same chats across the whole organization in a single
19request stream. For per-user listing, use `created_at.*` filters (or
20no time filter) with the default `order_by`. `user_ids[]` with
21`order_by=updated_at` is already rejected.
22
1423### Query Parameters
1524
1625- `after_id: optional string`
from line 72
6372
6473 - `gt: optional string`
6574
66 Filter chats updated after this time (RFC 3339 format)
75 Filter chats updated after this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
6776
6877 - `gte: optional string`
6978
70 Filter chats updated at or after this time (RFC 3339 format)
79 Filter chats updated at or after this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7180
7281 - `lt: optional string`
7382
74 Filter chats updated before this time (RFC 3339 format)
83 Filter chats updated before this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7584
7685 - `lte: optional string`
7786
78 Filter chats updated at or before this time (RFC 3339 format)
87 Filter chats updated at or before this time (RFC 3339 format). Combining updated_at filters with `user_ids[]` is deprecated and will be rejected after 2026-09-22; for updated_at-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
7988
8089- `user_ids: optional array of string`
8190
82 Filter to chats created by specific users (max 10 per request). Omit for an org-wide query. Enumerate IDs via `GET /v1/compliance/organizations/{org_uuid}/users`.
91 Filter to chats created by specific users (max 10 per request). Omit for an org-wide query. Enumerate IDs via `GET /v1/compliance/organizations/{org_uuid}/users`. Deprecated combination: passing `user_ids[]` together with any `updated_at.*` filter is deprecated and will be rejected after 2026-09-22. For `updated_at`-windowed polling, omit `user_ids[]` and use `order_by=updated_at` with `after_id` pagination.
8392
8493### Header Parameters
8594
from line 118
109118
110119 - `model: string or null`
111120
112 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
121 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
113122
114123 - `name: string`
115124
from line 182
173182 "created_at": "2025-06-07T08:09:10Z",
174183 "updated_at": "2025-06-07T09:10:11Z",
175184 "organization_id": "org_abc123",
176 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
185 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
177186 "project_id": "claude_proj_xyz789",
178 "model": "claude-opus-4-7",
187 "model": "claude-opus-5",
179188 "user": {
180189 "id": "user_xyz456",
api/compliance/apps/chats/messages Changed · +3 / -3 lines
from line 311
311311
312312- `model: string or null`
313313
314 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
314 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
315315
316316- `name: string`
317317
from line 361
361361 "created_at": "2025-06-07T08:09:10Z",
362362 "updated_at": "2025-06-07T08:09:11Z",
363363 "organization_id": "org_abc123",
364 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
364 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
365365 "project_id": "claude_proj_xyz789",
366 "model": "claude-opus-4-7",
366 "model": "claude-opus-5",
367367 "user": {
368368 "id": "user_xyz456",
api/compliance/apps/chats/messages/list Changed · +3 / -3 lines
from line 309
309309
310310- `model: string or null`
311311
312 Model selected for this chat (e.g. 'claude-opus-4-7'). May be null for legacy chats that never had a model recorded.
312 Model selected for this chat (e.g. 'claude-opus-5'). May be null for legacy chats that never had a model recorded.
313313
314314- `name: string`
315315
from line 359
359359 "created_at": "2025-06-07T08:09:10Z",
360360 "updated_at": "2025-06-07T08:09:11Z",
361361 "organization_id": "org_abc123",
362 "organization_uuid": "abcdef0123-4567-89ab-cdef-0123456789ab",
362 "organization_uuid": "abcdef01-2345-6789-abcd-ef0123456789",
363363 "project_id": "claude_proj_xyz789",
364 "model": "claude-opus-4-7",
364 "model": "claude-opus-5",
365365 "user": {
366366 "id": "user_xyz456",
api/compliance/apps/sessions Changed · +52 / -14 lines
from line 36
3636
3737 Opaque pagination token from a previous response's `next_page` field. Pass this to retrieve the next page of results. Clients should treat this value as an opaque string and not attempt to parse or interpret its contents, as the format may change without notice.
3838
39- `updated_at: optional object { gte }`
40
41 - `gte: optional string`
42
43 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.
44
3945### Header Parameters
4046
4147- `"x-api-key": optional string`
from line 48
4248
4349### Returns
4450
45- `data: array of object { id, created_at, organization_uuid, 4 more }`
51- `data: array of object { id, created_at, organization_uuid, 5 more }`
4652
47 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order.
53 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order. `updated_at` never participates in the ordering; the `updated_at.gte` query parameter filters on it without changing the order or the pagination cursor.
4854
4955 - `id: string`
5056
from line 72
6672
6773 - `"compliance_local_session"`
6874
75 - `updated_at: string`
76
77 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.
78
6979 - `user: object { id, email_address }`
7080
7181 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 118
109119 },
110120 "product_surface": "cowork",
111 "created_at": "2026-07-09T14:02:11Z"
121 "created_at": "2026-07-09T14:02:11Z",
122 "updated_at": "2026-07-09T15:47:33Z"
112123 }
113124 ]
114125}
from line 166
155166
156167 - `"compliance_local_session"`
157168
169- `updated_at: string`
170
171 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.
172
158173- `user: object { id, email_address }`
159174
160175 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 206
192207 },
193208 "product_surface": "cowork",
194 "created_at": "2026-07-09T14:02:11Z"
209 "created_at": "2026-07-09T14:02:11Z",
210 "updated_at": "2026-07-09T15:47:33Z"
195211}
196212```
197213
from line 215
199215
200216### Local List Response
201217
202- `LocalListResponse object { id, created_at, organization_uuid, 4 more }`
218- `LocalListResponse object { id, created_at, organization_uuid, 5 more }`
203219
204220 A Cowork or Claude Code session that a user ran on their own computer
205221 while signed in with their organization account.
from line 240
224240
225241 - `"compliance_local_session"`
226242
243 - `updated_at: string`
244
245 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.
246
227247 - `user: object { id, email_address }`
228248
229249 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 262
242262
243263### Local Retrieve Response
244264
245- `LocalRetrieveResponse object { id, created_at, organization_uuid, 4 more }`
265- `LocalRetrieveResponse object { id, created_at, organization_uuid, 5 more }`
246266
247267 A Cowork or Claude Code session that a user ran on their own computer
248268 while signed in with their organization account.
from line 287
267287
268288 - `"compliance_local_session"`
269289
290 - `updated_at: string`
291
292 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.
293
270294 - `user: object { id, email_address }`
271295
272296 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 359
335359
336360### Returns
337361
338- `data: array of object { id, content, created_at, 3 more }`
362- `data: array of object { id, content, created_at, 4 more }`
339363
340 Transcript turns for this page, ordered by `created_at` in the direction selected by the `order` parameter (ascending by default). Turns sharing a `created_at` (all messages of one inference call carry the call's timestamp) are returned in transcript order.
364 Transcript turns for this page, in call order: oldest call first by default, newest call first with `order=desc`. The messages of one call carry the call's timestamp and follow each other in transcript order; a page boundary can fall between them.
341365
342366 - `id: string`
343367
from line 451
427451
428452 When the message was recorded (RFC 3339, UTC)
429453
454 - `model: string or null`
455
456 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
457
430458 - `provenance: object { reason, type } or object { type } or object { type } or null`
431459
432460 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 465
437465
438466 - `reason: string`
439467
440 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
468 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
441469
442470 - `type: "content_unavailable"`
443471
from line 515
487515
488516 Opaque pagination cursor (prefixed `page_`) for the next page. Null when there is no further page. Treat as an opaque string; the format may change without notice.
489517
490- `session: object { id, created_at, organization_uuid, 4 more }`
518- `session: object { id, created_at, organization_uuid, 5 more }`
491519
492520 The local session the messages belong to. `user.email_address` is always null on this endpoint; the messages endpoint does not resolve email addresses.
493521
from line 539
511539
512540 - `"compliance_local_session"`
513541
542 - `updated_at: string`
543
544 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.
545
514546 - `user: object { id, email_address }`
515547
516548 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 581
549581 }
550582 ],
551583 "created_at": "2025-03-12T18:22:41.123456Z",
584 "model": "claude-opus-5",
552585 "provenance": {
553586 "reason": "not_captured",
554587 "type": "content_unavailable"
from line 597
564597 "organization_uuid": "a1b2c3d4-e5f6-4789-a012-3456789abcde",
565598 "product_surface": "cowork",
566599 "type": "compliance_local_session",
600 "updated_at": "2025-03-12T18:22:41.123456Z",
567601 "user": {
568602 "id": "user_01WCz1FkmYMm4gnmykNKUu3Q",
from line 611
577611
578612### Message List Response
579613
580- `MessageListResponse object { id, content, created_at, 3 more }`
614- `MessageListResponse object { id, content, created_at, 4 more }`
581615
582616 A single user or assistant turn in a local session transcript.
583617
from line 703
669703
670704 When the message was recorded (RFC 3339, UTC)
671705
706 - `model: string or null`
707
708 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
709
672710 - `provenance: object { reason, type } or object { type } or object { type } or null`
673711
674712 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 717
679717
680718 - `reason: string`
681719
682 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
720 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
683721
684722 - `type: "content_unavailable"`
685723
api/compliance/apps/sessions/local Changed · +52 / -14 lines
from line 34
3434
3535 Opaque pagination token from a previous response's `next_page` field. Pass this to retrieve the next page of results. Clients should treat this value as an opaque string and not attempt to parse or interpret its contents, as the format may change without notice.
3636
37- `updated_at: optional object { gte }`
38
39 - `gte: optional string`
40
41 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.
42
3743### Header Parameters
3844
3945- `"x-api-key": optional string`
from line 46
4046
4147### Returns
4248
43- `data: array of object { id, created_at, organization_uuid, 4 more }`
49- `data: array of object { id, created_at, organization_uuid, 5 more }`
4450
45 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order.
51 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order. `updated_at` never participates in the ordering; the `updated_at.gte` query parameter filters on it without changing the order or the pagination cursor.
4652
4753 - `id: string`
4854
from line 70
6470
6571 - `"compliance_local_session"`
6672
73 - `updated_at: string`
74
75 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.
76
6777 - `user: object { id, email_address }`
6878
6979 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 116
107117 },
108118 "product_surface": "cowork",
109 "created_at": "2026-07-09T14:02:11Z"
119 "created_at": "2026-07-09T14:02:11Z",
120 "updated_at": "2026-07-09T15:47:33Z"
110121 }
111122 ]
112123}
from line 164
153164
154165 - `"compliance_local_session"`
155166
167- `updated_at: string`
168
169 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.
170
156171- `user: object { id, email_address }`
157172
158173 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 204
190205 },
191206 "product_surface": "cowork",
192 "created_at": "2026-07-09T14:02:11Z"
207 "created_at": "2026-07-09T14:02:11Z",
208 "updated_at": "2026-07-09T15:47:33Z"
193209}
194210```
195211
from line 213
197213
198214### Local List Response
199215
200- `LocalListResponse object { id, created_at, organization_uuid, 4 more }`
216- `LocalListResponse object { id, created_at, organization_uuid, 5 more }`
201217
202218 A Cowork or Claude Code session that a user ran on their own computer
203219 while signed in with their organization account.
from line 238
222238
223239 - `"compliance_local_session"`
224240
241 - `updated_at: string`
242
243 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.
244
225245 - `user: object { id, email_address }`
226246
227247 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 260
240260
241261### Local Retrieve Response
242262
243- `LocalRetrieveResponse object { id, created_at, organization_uuid, 4 more }`
263- `LocalRetrieveResponse object { id, created_at, organization_uuid, 5 more }`
244264
245265 A Cowork or Claude Code session that a user ran on their own computer
246266 while signed in with their organization account.
from line 285
265285
266286 - `"compliance_local_session"`
267287
288 - `updated_at: string`
289
290 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.
291
268292 - `user: object { id, email_address }`
269293
270294 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 357
333357
334358### Returns
335359
336- `data: array of object { id, content, created_at, 3 more }`
360- `data: array of object { id, content, created_at, 4 more }`
337361
338 Transcript turns for this page, ordered by `created_at` in the direction selected by the `order` parameter (ascending by default). Turns sharing a `created_at` (all messages of one inference call carry the call's timestamp) are returned in transcript order.
362 Transcript turns for this page, in call order: oldest call first by default, newest call first with `order=desc`. The messages of one call carry the call's timestamp and follow each other in transcript order; a page boundary can fall between them.
339363
340364 - `id: string`
341365
from line 449
425449
426450 When the message was recorded (RFC 3339, UTC)
427451
452 - `model: string or null`
453
454 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
455
428456 - `provenance: object { reason, type } or object { type } or object { type } or null`
429457
430458 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 463
435463
436464 - `reason: string`
437465
438 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
466 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
439467
440468 - `type: "content_unavailable"`
441469
from line 513
485513
486514 Opaque pagination cursor (prefixed `page_`) for the next page. Null when there is no further page. Treat as an opaque string; the format may change without notice.
487515
488- `session: object { id, created_at, organization_uuid, 4 more }`
516- `session: object { id, created_at, organization_uuid, 5 more }`
489517
490518 The local session the messages belong to. `user.email_address` is always null on this endpoint; the messages endpoint does not resolve email addresses.
491519
from line 537
509537
510538 - `"compliance_local_session"`
511539
540 - `updated_at: string`
541
542 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.
543
512544 - `user: object { id, email_address }`
513545
514546 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 579
547579 }
548580 ],
549581 "created_at": "2025-03-12T18:22:41.123456Z",
582 "model": "claude-opus-5",
550583 "provenance": {
551584 "reason": "not_captured",
552585 "type": "content_unavailable"
from line 595
562595 "organization_uuid": "a1b2c3d4-e5f6-4789-a012-3456789abcde",
563596 "product_surface": "cowork",
564597 "type": "compliance_local_session",
598 "updated_at": "2025-03-12T18:22:41.123456Z",
565599 "user": {
566600 "id": "user_01WCz1FkmYMm4gnmykNKUu3Q",
from line 609
575609
576610### Message List Response
577611
578- `MessageListResponse object { id, content, created_at, 3 more }`
612- `MessageListResponse object { id, content, created_at, 4 more }`
579613
580614 A single user or assistant turn in a local session transcript.
581615
from line 701
667701
668702 When the message was recorded (RFC 3339, UTC)
669703
704 - `model: string or null`
705
706 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
707
670708 - `provenance: object { reason, type } or object { type } or object { type } or null`
671709
672710 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 715
677715
678716 - `reason: string`
679717
680 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
718 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
681719
682720 - `type: "content_unavailable"`
683721
api/compliance/apps/sessions/local/list Changed · +14 / -3 lines
from line 32
3232
3333 Opaque pagination token from a previous response's `next_page` field. Pass this to retrieve the next page of results. Clients should treat this value as an opaque string and not attempt to parse or interpret its contents, as the format may change without notice.
3434
35- `updated_at: optional object { gte }`
36
37 - `gte: optional string`
38
39 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.
40
3541### Header Parameters
3642
3743- `"x-api-key": optional string`
from line 44
3844
3945### Returns
4046
41- `data: array of object { id, created_at, organization_uuid, 4 more }`
47- `data: array of object { id, created_at, organization_uuid, 5 more }`
4248
43 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order.
49 Page of local sessions, ordered by `created_at` descending; ties are broken by a fixed server-side order. `updated_at` never participates in the ordering; the `updated_at.gte` query parameter filters on it without changing the order or the pagination cursor.
4450
4551 - `id: string`
4652
from line 68
6268
6369 - `"compliance_local_session"`
6470
71 - `updated_at: string`
72
73 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.
74
6575 - `user: object { id, email_address }`
6676
6777 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 114
105115 },
106116 "product_surface": "cowork",
107 "created_at": "2026-07-09T14:02:11Z"
117 "created_at": "2026-07-09T14:02:11Z",
118 "updated_at": "2026-07-09T15:47:33Z"
108119 }
109120 ]
110121}
api/compliance/apps/sessions/local/messages Changed · +21 / -7 lines
from line 55
5555
5656### Returns
5757
58- `data: array of object { id, content, created_at, 3 more }`
58- `data: array of object { id, content, created_at, 4 more }`
5959
60 Transcript turns for this page, ordered by `created_at` in the direction selected by the `order` parameter (ascending by default). Turns sharing a `created_at` (all messages of one inference call carry the call's timestamp) are returned in transcript order.
60 Transcript turns for this page, in call order: oldest call first by default, newest call first with `order=desc`. The messages of one call carry the call's timestamp and follow each other in transcript order; a page boundary can fall between them.
6161
6262 - `id: string`
6363
from line 147
147147
148148 When the message was recorded (RFC 3339, UTC)
149149
150 - `model: string or null`
151
152 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
153
150154 - `provenance: object { reason, type } or object { type } or object { type } or null`
151155
152156 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 161
157161
158162 - `reason: string`
159163
160 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
164 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
161165
162166 - `type: "content_unavailable"`
163167
from line 211
207211
208212 Opaque pagination cursor (prefixed `page_`) for the next page. Null when there is no further page. Treat as an opaque string; the format may change without notice.
209213
210- `session: object { id, created_at, organization_uuid, 4 more }`
214- `session: object { id, created_at, organization_uuid, 5 more }`
211215
212216 The local session the messages belong to. `user.email_address` is always null on this endpoint; the messages endpoint does not resolve email addresses.
213217
from line 235
231235
232236 - `"compliance_local_session"`
233237
238 - `updated_at: string`
239
240 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.
241
234242 - `user: object { id, email_address }`
235243
236244 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 277
269277 }
270278 ],
271279 "created_at": "2025-03-12T18:22:41.123456Z",
280 "model": "claude-opus-5",
272281 "provenance": {
273282 "reason": "not_captured",
274283 "type": "content_unavailable"
from line 293
284293 "organization_uuid": "a1b2c3d4-e5f6-4789-a012-3456789abcde",
285294 "product_surface": "cowork",
286295 "type": "compliance_local_session",
296 "updated_at": "2025-03-12T18:22:41.123456Z",
287297 "user": {
288298 "id": "user_01WCz1FkmYMm4gnmykNKUu3Q",
from line 307
297307
298308### Message List Response
299309
300- `MessageListResponse object { id, content, created_at, 3 more }`
310- `MessageListResponse object { id, content, created_at, 4 more }`
301311
302312 A single user or assistant turn in a local session transcript.
303313
from line 399
389399
390400 When the message was recorded (RFC 3339, UTC)
391401
402 - `model: string or null`
403
404 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
405
392406 - `provenance: object { reason, type } or object { type } or object { type } or null`
393407
394408 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 413
399413
400414 - `reason: string`
401415
402 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
416 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
403417
404418 - `type: "content_unavailable"`
405419
api/compliance/apps/sessions/local/messages/list Changed · +15 / -5 lines
from line 53
5353
5454### Returns
5555
56- `data: array of object { id, content, created_at, 3 more }`
56- `data: array of object { id, content, created_at, 4 more }`
5757
58 Transcript turns for this page, ordered by `created_at` in the direction selected by the `order` parameter (ascending by default). Turns sharing a `created_at` (all messages of one inference call carry the call's timestamp) are returned in transcript order.
58 Transcript turns for this page, in call order: oldest call first by default, newest call first with `order=desc`. The messages of one call carry the call's timestamp and follow each other in transcript order; a page boundary can fall between them.
5959
6060 - `id: string`
6161
from line 145
145145
146146 When the message was recorded (RFC 3339, UTC)
147147
148 - `model: string or null`
149
150 The model that served this assistant turn, as reported in the `model` field of the underlying Messages API response. Null on user messages and on any assistant message whose `provenance` is set: client-asserted history and synthetic markers were not produced by a model during this session, and for unavailable content the serving model is not known.
151
148152 - `provenance: object { reason, type } or object { type } or object { type } or null`
149153
150154 Where this turn's content came from, discriminated on `type`. Null (the common case) means verified content: on an assistant message, content Claude produced during this session; on a user message, content the user sent. `content_unavailable`: the turn's content cannot be returned and `content` is empty; `reason` says why. `client_asserted`: assistant content the client supplied as conversation history; `content` shows what the model received but its authorship is not verified; never on user-role messages. `synthetic_marker`: a transcript marker the endpoint generated rather than content either party sent during the session. Both `client_asserted` and `synthetic_marker` can result from normal request or client processing, not only client modification. Callers should tolerate unrecognized `type` values.
from line 159
155159
156160 - `reason: string`
157161
158 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
162 Why this turn's content cannot be returned, e.g. `not_captured` (the content was not captured for compliance retrieval), `client_aborted` (the client closed the connection or cancelled the request before the response completed, so the response was not captured for this turn; any partial output already streamed to the client is not included; assistant-role turns only), `cmek_key_revoked` (the content is encrypted under the organization's customer-managed key and that key is unavailable), `retention_elapsed` (the content lies past the organization's retention boundary; on the placeholder standing in for every pre-boundary turn), or `oversize` (the message exceeds the server's per-message size bound even after per-block truncation). Callers should tolerate unrecognized values. `not_captured` is not proof that no record was stored: content withheld by the storage layer's fail-closed access policies carries the same reason and is deliberately indistinguishable from content that was never captured.
159163
160164 - `type: "content_unavailable"`
161165
from line 209
205209
206210 Opaque pagination cursor (prefixed `page_`) for the next page. Null when there is no further page. Treat as an opaque string; the format may change without notice.
207211
208- `session: object { id, created_at, organization_uuid, 4 more }`
212- `session: object { id, created_at, organization_uuid, 5 more }`
209213
210214 The local session the messages belong to. `user.email_address` is always null on this endpoint; the messages endpoint does not resolve email addresses.
211215
from line 233
229233
230234 - `"compliance_local_session"`
231235
236 - `updated_at: string`
237
238 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.
239
232240 - `user: object { id, email_address }`
233241
234242 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 275
267275 }
268276 ],
269277 "created_at": "2025-03-12T18:22:41.123456Z",
278 "model": "claude-opus-5",
270279 "provenance": {
271280 "reason": "not_captured",
272281 "type": "content_unavailable"
from line 291
282291 "organization_uuid": "a1b2c3d4-e5f6-4789-a012-3456789abcde",
283292 "product_surface": "cowork",
284293 "type": "compliance_local_session",
294 "updated_at": "2025-03-12T18:22:41.123456Z",
285295 "user": {
286296 "id": "user_01WCz1FkmYMm4gnmykNKUu3Q",
api/compliance/apps/sessions/local/retrieve Changed · +6 / -1 lines
from line 44
4444
4545 - `"compliance_local_session"`
4646
47- `updated_at: string`
48
49 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.
50
4751- `user: object { id, email_address }`
4852
4953 The authenticated user at the time of the session. Always set; `user.id` is always populated. `user.email_address` is null when the user's account has been deleted or the user is no longer a member of an organization the key may read.
from line 84
8185 },
8286 "product_surface": "cowork",
83 "created_at": "2026-07-09T14:02:11Z"
87 "created_at": "2026-07-09T14:02:11Z",
88 "updated_at": "2026-07-09T15:47:33Z"
8489}
8590```
8691
api/compliance/organizations Changed · +6 / -2 lines
from line 625
625625
626626 A setting whose enforced value is a single true/false flag.
627627
628 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 43 more`
628 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 44 more`
629629
630630 - `"ai_powered_artifacts_enabled"`
631631
from line 685
685685
686686 - `"frontier_data_use_enabled"`
687687
688 - `"group_skill_sharing_enabled"`
689
688690 - `"hipaa_compliance_enabled"`
689691
690692 - `"inline_visualizations_enabled"`
from line 950
948950
949951 A setting whose enforced value is a single true/false flag.
950952
951 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 43 more`
953 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 44 more`
952954
953955 - `"ai_powered_artifacts_enabled"`
954956
from line 1009
10071009 - `"directory_sync_enabled"`
10081010
10091011 - `"frontier_data_use_enabled"`
1012
1013 - `"group_skill_sharing_enabled"`
10101014
10111015 - `"hipaa_compliance_enabled"`
10121016
api/compliance/organizations/settings Changed · +6 / -2 lines
from line 76
7676
7777 A setting whose enforced value is a single true/false flag.
7878
79 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 43 more`
79 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 44 more`
8080
8181 - `"ai_powered_artifacts_enabled"`
8282
from line 136
136136
137137 - `"frontier_data_use_enabled"`
138138
139 - `"group_skill_sharing_enabled"`
140
139141 - `"hipaa_compliance_enabled"`
140142
141143 - `"inline_visualizations_enabled"`
from line 401
399401
400402 A setting whose enforced value is a single true/false flag.
401403
402 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 43 more`
404 - `name: "ai_powered_artifacts_enabled" or "api_workbench_feedback_collection_enabled" or "artifact_connectors_enabled" or 44 more`
403405
404406 - `"ai_powered_artifacts_enabled"`
405407
from line 460
458460 - `"directory_sync_enabled"`
459461
460462 - `"frontier_data_use_enabled"`
463
464 - `"group_skill_sharing_enabled"`
461465
462466 - `"hipaa_compliance_enabled"`
463467
home Changed · +8 / -8 lines
from line 86
8686 Get API key
8787 </HomeJourneyLink>
8888
89 <HomeJourneyLink icon="settings" href="https://platform.claude.com/docs/en/about-claude/models/overview">
89 <HomeJourneyLink icon="settings" href="https://platform.claude.com/docs/en/models/overview">
9090 Choose a model
9191 </HomeJourneyLink>
9292
from line 242
242242 <HomeSectionHeader label="Models" title="The Claude model family" description="Choose the right model for your use case." />
243243
244244 <HomeModelCards>
245 <HomeModelCard name="Fable 5" badge="Most capable" modelId="claude-fable-5" description="Highest capability for the most demanding reasoning and long-horizon agentic work." href="https://platform.claude.com/docs/en/about-claude/models/overview" />
245 <HomeModelCard name="Fable 5" badge="Most capable" modelId="claude-fable-5" description="Highest capability for the most demanding reasoning and long-horizon agentic work." href="https://platform.claude.com/docs/en/models/fable-5/overview" />
246246
247 <HomeModelCard name="Opus 5" badge="Advanced" modelId="claude-opus-5" description="Excellent for complex analysis, coding, and creative tasks requiring deep reasoning." href="https://platform.claude.com/docs/en/about-claude/models/overview" />
247 <HomeModelCard name="Opus 5" badge="Advanced" modelId="claude-opus-5" description="Excellent for complex analysis, coding, and creative tasks requiring deep reasoning." href="https://platform.claude.com/docs/en/models/opus-5/overview" />
248248
249 <HomeModelCard name="Sonnet 5" badge="Best balance" modelId="claude-sonnet-5" description="Ideal balance of intelligence and speed for most production workloads." href="https://platform.claude.com/docs/en/about-claude/models/overview" />
249 <HomeModelCard name="Sonnet 5" badge="Best balance" modelId="claude-sonnet-5" description="Ideal balance of intelligence and speed for most production workloads." href="https://platform.claude.com/docs/en/models/sonnet-5/overview" />
250250
251 <HomeModelCard name="Haiku 4.5" badge="Fastest" modelId="claude-haiku-4-5" description="Lightning-fast responses for high-volume, latency-sensitive applications." href="https://platform.claude.com/docs/en/about-claude/models/overview" />
251 <HomeModelCard name="Haiku 4.5" badge="Fastest" modelId="claude-haiku-4-5" description="Lightning-fast responses for high-volume, latency-sensitive applications." href="https://platform.claude.com/docs/en/models/haiku-4-5/overview" />
252252 </HomeModelCards>
253253 </HomeSection>
254254
manage-claude/compliance-faq Changed · +16 / -29 lines
from line 24
2424 </Accordion>
2525
2626 <Accordion title="Is turning the Compliance API off in Claude Console recorded anywhere?">
27 Yes. When the Compliance API is turned off (or back on) in Claude Console, the change is recorded as an organization settings-updated activity in the [Activity Feed](https://platform.claude.com/docs/en/manage-claude/compliance-activity-feed), so your audit trail shows who changed the setting and when. This activity is an exception to the recording stop: the disable is recorded even though no other activity is recorded while the Compliance API is off.
27 Yes. When the Compliance API is turned off (or back on) in Claude Console, the change is recorded as an `org_compliance_api_settings_updated` activity in the [Activity Feed](https://platform.claude.com/docs/en/manage-claude/compliance-activity-feed), so your audit trail shows who changed the setting and when. This activity is an exception to the recording stop: the disable is recorded even though no other activity is recorded while the Compliance API is off.
2828 </Accordion>
2929
3030 <Accordion title="Why doesn't my parent organization appear in Claude Console when creating an Admin API key?">
from line 52
5252 }
5353 ```
5454
55 To access content endpoints, the primary owner of your parent organization (or an organization owner, for their own organization only) must [create a Compliance Access Key](https://platform.claude.com/docs/en/manage-claude/compliance-api-access#set-up-the-compliance-api) with `read:compliance_user_data` (and `delete:compliance_user_data` for deletes), or `read:compliance_org_data` for organization, role, group, and effective-settings endpoints. See [Handle Compliance API errors](https://platform.claude.com/docs/en/manage-claude/compliance-errors#403-forbidden) for the full per-endpoint catalog.
55 To access content endpoints, the primary owner of your parent organization (or an organization owner, for their own organization only) must [create a Compliance Access Key](https://platform.claude.com/docs/en/manage-claude/compliance-api-access#set-up-the-compliance-api) with `read:compliance_user_data` (and `delete:compliance_user_data` for deletes), or `read:compliance_org_data` for organization, role, group, and effective-settings endpoints. A standalone Claude Console organization (one with no parent organization) cannot create a Compliance Access Key, so the content endpoints are not available to it; it can query the Activity Feed only. See [Handle Compliance API errors](https://platform.claude.com/docs/en/manage-claude/compliance-errors#403-forbidden) for the full per-endpoint catalog.
5656 </Accordion>
5757</AccordionGroup>
5858
from line 60
6060
6161<AccordionGroup>
6262 <Accordion title="How far back does the Activity Feed go?">
63 The Activity Feed retains 6 years of organization activity, and new events are queryable within 1 minute of occurring. Activity Feed retention is independent of your organization's content retention policy: chat, file, and project content follows the retention rules configured for your organization (indefinite by default).
63 The Activity Feed retains 6 years of organization activity, and new events are queryable within 1 minute of occurring. The feed reaches back at most to the point the Compliance API was first enabled for your organization: recording is not retroactive, and activity from before enablement is not backfilled. Activity Feed retention is independent of your organization's content retention policy: chat, file, and project content follows the retention rules configured for your organization (indefinite by default).
6464 </Accordion>
6565
6666 <Accordion title="Does the Activity Feed include prompt or message content?">
from line 80
8080 <Accordion title="What do Cowork and Claude Code session transcripts include?">
8181 Local and remote session transcripts both carry user prompts, assistant responses, and tool calls and results. For local sessions (Cowork and Claude Code on users' machines), that is what Claude was asked to do and what it returned, not what happened on the device.
8282
83 | Data | Local sessions (on users' machines) | Remote sessions (in the cloud) |
84 | --------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
85 | User prompts | Yes; returned as `text` blocks. | Yes; returned as `text` blocks. |
86 | Assistant responses | Yes; text output only. | Yes; text output only. |
87 | Tool calls and results | Yes; each `tool_use` input and each `text` entry in a `tool_result` is truncated to 10,000 bytes by default (up to about 1 MiB each on request). | Yes; each `tool_use` input and each `text` entry in a `tool_result` is truncated to 10,000 bytes by default (up to about 1 MiB each on request). |
88 | File contents and file names | Yes; text that Claude reads through tools appears in the transcript, subject to the same truncation. Images, PDFs, and other binary or structured content appear only as placeholder `text` blocks. File names appear in tool-call inputs and outputs. | Yes; file contents and file names appear in the transcript through tool-call inputs and outputs (text only; other content is omitted). |
89 | Artifacts | Yes; generated content appears inside tool-call inputs in the transcript. | Yes; generated content appears inside tool-call inputs in the transcript. |
90 | Skills | Yes; skill content appears when the client sends it as message content, and it is not distinguished from other user text. | Yes; skill content appears in the transcript. |
91 | Session metadata | Yes; owner (`user.id` and email address), organization, workspace, `product_surface`, and `created_at`, from the list and retrieve endpoints. Local sessions carry no `status` or `updated_at`. | Yes; owner, organization, status, timestamps, and `product_surface`, from the list endpoint. |
92 | Thinking blocks | No. | No. |
93 | Images and other non-text content | No; each image, PDF, or other binary or structured block appears as a placeholder `text` block (for example, `[image content not shown]`) with `truncated` set to `true`. Raw file bytes are never returned. | No; non-text blocks are omitted, and raw file bytes are never returned. |
94 | Token usage, cost, and latency | No; use [Cowork's OpenTelemetry logging](https://support.claude.com/en/articles/14477985-monitor-claude-cowork-activity-with-opentelemetry) or [Claude Code monitoring](https://code.claude.com/docs/en/monitoring-usage) for usage and performance telemetry. | No; use [OpenTelemetry logging](https://support.claude.com/en/articles/14477985-monitor-claude-cowork-activity-with-opentelemetry) for usage and performance telemetry. |
83 | Data | Local sessions (on users' machines) | Remote sessions (in the cloud) |
84 | --------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
85 | User prompts | Yes; returned as `text` blocks. | Yes; returned as `text` blocks. |
86 | Assistant responses | Yes; text output only. | Yes; text output only. |
87 | Tool calls and results | Yes; each `tool_use` input and each `text` entry in a `tool_result` is truncated to 10,000 bytes by default (up to about 1 MiB each on request). | Yes; each `tool_use` input and each `text` entry in a `tool_result` is truncated to 10,000 bytes by default (up to about 1 MiB each on request). |
88 | File contents and file names | Yes; text that Claude reads through tools appears in the transcript, subject to the same truncation. Images, PDFs, and other binary or structured content appear only as placeholder `text` blocks. File names appear in tool-call inputs and outputs. | Yes; file contents and file names appear in the transcript through tool-call inputs and outputs (text only; other content is omitted). |
89 | Artifacts | Yes; generated content appears inside tool-call inputs in the transcript. | Yes; generated content appears inside tool-call inputs in the transcript. |
90 | Skills | Yes; skill content appears when the client sends it as message content, and it is not distinguished from other user text. | Yes; skill content appears in the transcript. |
91 | Session metadata | Yes; owner (`user.id` and email address), organization, workspace, `product_surface`, `created_at`, and `updated_at`, from the list and retrieve endpoints. Local sessions carry no `status`. | Yes; owner, organization, status, timestamps, and `product_surface`, from the list endpoint. |
92 | Thinking blocks | No. | No. |
93 | Images and other non-text content | No; each image, PDF, or other binary or structured block appears as a placeholder `text` block (for example, `[image content not shown]`) with `truncated` set to `true`. Raw file bytes are never returned. | No; non-text blocks are omitted, and raw file bytes are never returned. |
94 | Token usage, cost, and latency | No; token usage and cost are available through the [Claude Enterprise Analytics API](https://platform.claude.com/docs/en/manage-claude/analytics-api#get-access-to-the-claude-enterprise-analytics-api). | No; token usage and cost are available through the [Claude Enterprise Analytics API](https://platform.claude.com/docs/en/manage-claude/analytics-api#get-access-to-the-claude-enterprise-analytics-api). |
9595
9696 See [Sessions on users' machines](https://platform.claude.com/docs/en/manage-claude/compliance-sessions#retrieve-local-sessions) and [Sessions in the cloud](https://platform.claude.com/docs/en/manage-claude/compliance-sessions#retrieve-remote-sessions) for the endpoints and parameters.
9797 </Accordion>
9898
9999 <Accordion title="How does session coverage compare with OpenTelemetry logging (OTEL) for Cowork and Claude Code?">
100 [Cowork's OpenTelemetry logging](https://support.claude.com/en/articles/14477985-monitor-claude-cowork-activity-with-opentelemetry) and [Claude Code monitoring](https://code.claude.com/docs/en/monitoring-usage) overlap with the session endpoints but answer different needs: OTEL streams per-event telemetry to infrastructure you run as activity happens, whereas the Compliance API lets you retrieve retained per-session transcripts from Anthropic after the fact.
101
102 | | Local sessions (on users' machines) | Remote sessions (in the cloud) | OpenTelemetry logging |
103 | --------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- |
104 | Delivery | Pull: query and export over HTTPS | Pull: query and export over HTTPS | Push: streamed to your OTLP collector |
105 | Setup | Works with your existing Compliance Access Key | Works with your existing Compliance Access Key | Admin configures an OTLP endpoint and content-capture settings |
106 | Infrastructure | Anthropic-hosted | Anthropic-hosted | You run the collector and storage |
107 | Retention | 6 years by default, or your organization's custom conversation retention period when a finite one is set; held by Anthropic | 6 years, held by Anthropic | Your infrastructure, your policies |
108 | User prompts and assistant responses | Yes | Yes | Yes, subject to content-capture settings |
109 | Tool inputs | Truncated to 10,000 bytes per input by default; up to about 1 MiB on request | Truncated to 10,000 bytes per input by default; up to about 1 MiB on request | Truncated summaries |
110 | Tool result content | Each text entry truncated to 10,000 bytes by default; up to about 1 MiB on request | Each text entry truncated to 10,000 bytes by default; up to about 1 MiB on request | Metadata such as size and success; Claude Code can also capture content with an optional, size-capped setting |
111 | File contents | Yes, through transcript tool calls (text only; other content appears as a placeholder) | Yes, through transcript tool calls (text only; other content is omitted) | File paths; Claude Code can also capture contents with an optional, size-capped setting |
112 | Host and device metadata (terminal type, workspace paths) | No | No | Yes |
113 | Token usage and cost | No | No | Yes |
100 [Cowork's OpenTelemetry logging](https://support.claude.com/en/articles/14477985-monitor-claude-cowork-activity-with-opentelemetry) and [Claude Code monitoring](https://code.claude.com/docs/en/monitoring-usage) overlap with the session endpoints but answer different needs: OTEL streams per-event telemetry to infrastructure you run as activity happens, whereas the Compliance API lets you retrieve retained per-session transcripts from Anthropic after the fact. OTEL can also capture prompts and responses, but Anthropic recommends the Compliance API for retrieving the content of Cowork and Claude Code sessions. For a table comparing local sessions, remote sessions, and OTEL, see the introduction to [Retrieve session transcripts](https://platform.claude.com/docs/en/manage-claude/compliance-sessions).
114101
115102 OTEL events and Compliance API records share organization and user identifiers, so you can join them.
116103 </Accordion>