chore: update deepseek provider docs - #5177
Conversation
📝 WalkthroughWalkthroughDeepSeek provider documentation distinguishes Anthropic-compatible Chat Completions and Responses endpoints from the OpenAI-compatible Text Completions endpoint, updating routing, authentication, parameter mapping, request handling, and compatibility caveats. ChangesDeepSeek compatibility documentation
Estimated code review effort: 2 (Simple) | ~10 minutes Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
This stack of pull requests is managed by Graphite. Learn more about stacking. |
cda389d to
46eed16
Compare
c5a5dad to
c0d7ee2
Compare
Confidence Score: 5/5This looks safe to merge.
Important Files Changed
Reviews (14): Last reviewed commit: "chore: update deepseek provider docs" | Re-trigger Greptile |
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/providers/supported-providers/deepseek.mdx`:
- Around line 195-199: Update the “Impact” statement in the “Mixed Endpoint
Compatibility” accordion to limit the OpenAI caveat to endpoint, authentication,
and request/response conversion details. Preserve that shared capabilities such
as reasoning, tools, streaming, and extra_params remain applicable, and retain
the Anthropic documentation cross-reference for Chat Completions and Responses.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: caad0b40-cc4c-4a53-b1a2-cbec9d1a6f13
📒 Files selected for processing (1)
docs/providers/supported-providers/deepseek.mdx
46eed16 to
215f151
Compare
c0d7ee2 to
39fa3f2
Compare
39fa3f2 to
2d5cfe3
Compare
215f151 to
7fe0b4e
Compare
2d5cfe3 to
7455a0a
Compare
7fe0b4e to
5692d79
Compare
7d7238a to
2eac7a9
Compare
5692d79 to
99ebf69
Compare
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/providers/supported-providers/deepseek.mdx`:
- Around line 114-116: Remove or relocate the “Extra Parameters” subsection from
the Chat Completions documentation so it is not attributed to that endpoint. If
retained, place it under the provider’s Text Completions operation, preserving
the documented passthrough behavior and endpoint-specific alignment.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 833fe6a6-3d1d-4302-ae0b-c64322c8a56f
📒 Files selected for processing (1)
docs/providers/supported-providers/deepseek.mdx
ce2d3c4 to
b7df125
Compare
2e6576d to
700566a
Compare
700566a to
609bfb6
Compare
b7df125 to
1422dfc
Compare
609bfb6 to
7d5a1a5
Compare
1422dfc to
68d4683
Compare
7d5a1a5 to
d1d1779
Compare
68d4683 to
491ce5e
Compare
491ce5e to
e6b494b
Compare
d1d1779 to
0717f57
Compare
Merge activity
|
The base branch was changed.
## Summary Updates the DeepSeek provider documentation to reflect that Chat Completions and the Responses API are now routed through DeepSeek's Anthropic-compatible endpoint (`/anthropic/v1/messages`) rather than the OpenAI-compatible endpoint. Text Completions continue to use the OpenAI-compatible `/beta/completions` beta endpoint. ## Changes - Updated the overview to describe DeepSeek as a mixed-endpoint provider: Anthropic-compatible for Chat Completions and Responses API, OpenAI-compatible for Text Completions - Replaced references to OpenAI-compatible chat behavior with Anthropic-compatible equivalents, including authentication (`x-api-key` instead of `Authorization: Bearer`), request/response converters, tool calling, and reasoning/thinking parameter mapping - Removed the documented internal conversion flow where Responses API requests were first converted to Chat Completions — Responses requests are now sent natively to `/anthropic/v1/messages` - Removed the filtered parameters section (previously OpenAI-specific exclusions) and the `stripReasoningDetails` behavior note, as these no longer apply to the Anthropic-compatible path - Updated the "Known Limitations" accordions to clarify that the user field size limit applies only to Text Completions, and replaced the "Reasoning Details in Assistant Messages" limitation with a "Mixed Endpoint Compatibility" note explaining the two-endpoint architecture and its implications for which provider docs apply ## Type of change - [ ] Bug fix - [ ] Feature - [ ] Refactor - [x] Documentation - [ ] Chore/CI ## Affected areas - [ ] Core (Go) - [ ] Transports (HTTP) - [ ] Providers/Integrations - [ ] Plugins - [ ] UI (React) - [x] Docs ## How to test Review the updated `docs/providers/supported-providers/deepseek.mdx` to confirm: - The overview accurately describes the Anthropic-compatible endpoint for Chat Completions and Responses API - Cross-references point to the Anthropic provider docs instead of OpenAI for chat and responses behavior - The Text Completions section still correctly references the OpenAI-compatible `/beta/completions` endpoint - The known limitations section reflects the current mixed-endpoint architecture ## Breaking changes - [ ] Yes - [x] No ## Security considerations The documentation now explicitly calls out that the Anthropic-compatible endpoint uses `x-api-key` header authentication rather than a bearer token, which is relevant for operators reviewing how provider credentials are transmitted. ## Checklist - [x] I read `docs/contributing/README.md` and followed the guidelines - [x] I added/updated tests where appropriate - [x] I updated documentation where needed - [x] I verified builds succeed (Go and UI) - [x] I verified the CI pipeline passes locally if applicable
## Summary Updates the DeepSeek provider documentation to reflect that Chat Completions and the Responses API are now routed through DeepSeek's Anthropic-compatible endpoint (`/anthropic/v1/messages`) rather than the OpenAI-compatible endpoint. Text Completions continue to use the OpenAI-compatible `/beta/completions` beta endpoint. ## Changes - Updated the overview to describe DeepSeek as a mixed-endpoint provider: Anthropic-compatible for Chat Completions and Responses API, OpenAI-compatible for Text Completions - Replaced references to OpenAI-compatible chat behavior with Anthropic-compatible equivalents, including authentication (`x-api-key` instead of `Authorization: Bearer`), request/response converters, tool calling, and reasoning/thinking parameter mapping - Removed the documented internal conversion flow where Responses API requests were first converted to Chat Completions — Responses requests are now sent natively to `/anthropic/v1/messages` - Removed the filtered parameters section (previously OpenAI-specific exclusions) and the `stripReasoningDetails` behavior note, as these no longer apply to the Anthropic-compatible path - Updated the "Known Limitations" accordions to clarify that the user field size limit applies only to Text Completions, and replaced the "Reasoning Details in Assistant Messages" limitation with a "Mixed Endpoint Compatibility" note explaining the two-endpoint architecture and its implications for which provider docs apply ## Type of change - [ ] Bug fix - [ ] Feature - [ ] Refactor - [x] Documentation - [ ] Chore/CI ## Affected areas - [ ] Core (Go) - [ ] Transports (HTTP) - [ ] Providers/Integrations - [ ] Plugins - [ ] UI (React) - [x] Docs ## How to test Review the updated `docs/providers/supported-providers/deepseek.mdx` to confirm: - The overview accurately describes the Anthropic-compatible endpoint for Chat Completions and Responses API - Cross-references point to the Anthropic provider docs instead of OpenAI for chat and responses behavior - The Text Completions section still correctly references the OpenAI-compatible `/beta/completions` endpoint - The known limitations section reflects the current mixed-endpoint architecture ## Breaking changes - [ ] Yes - [x] No ## Security considerations The documentation now explicitly calls out that the Anthropic-compatible endpoint uses `x-api-key` header authentication rather than a bearer token, which is relevant for operators reviewing how provider credentials are transmitted. ## Checklist - [x] I read `docs/contributing/README.md` and followed the guidelines - [x] I added/updated tests where appropriate - [x] I updated documentation where needed - [x] I verified builds succeed (Go and UI) - [x] I verified the CI pipeline passes locally if applicable

Summary
Updates the DeepSeek provider documentation to reflect that Chat Completions and the Responses API are now routed through DeepSeek's Anthropic-compatible endpoint (
/anthropic/v1/messages) rather than the OpenAI-compatible endpoint. Text Completions continue to use the OpenAI-compatible/beta/completionsbeta endpoint.Changes
x-api-keyinstead ofAuthorization: Bearer), request/response converters, tool calling, and reasoning/thinking parameter mapping/anthropic/v1/messagesstripReasoningDetailsbehavior note, as these no longer apply to the Anthropic-compatible pathType of change
Affected areas
How to test
Review the updated
docs/providers/supported-providers/deepseek.mdxto confirm:/beta/completionsendpointBreaking changes
Security considerations
The documentation now explicitly calls out that the Anthropic-compatible endpoint uses
x-api-keyheader authentication rather than a bearer token, which is relevant for operators reviewing how provider credentials are transmitted.Checklist
docs/contributing/README.mdand followed the guidelines