🐛 fix(api): Fix the issue where custom AI Providers cannot use custom APIs

[hardy-one]

💻 Change Type

• ✨ feat
• 🐛 fix
• ♻️ refactor
• 💄 style
• 👷 build
• ⚡️ perf
• ✅ test
• 📝 docs
• 🔨 chore

🔗 Related Issue

Fixes #11296 Fixes #11271

🔀 Description of Change

修复自定义 AI Provider 无法使用服务端 API 的问题。

问题原因：
客户端使用 sdkType（如 'azure'、'openai'）构建 API 端点路径，但服务端使用该路径作为 provider ID 从数据库查询配置。这导致服务端查询了内置 provider 的配置，而不是用户自定义 provider 的配置。

修复方案：
将 API 端点路径从使用 sdkType 改为使用原始的 provider 标识。这确保服务端能正确查询用户配置，然后从配置中读取 sdkType 来初始化相应的运行时。

修改文件：
src/services/chat/index.ts：聊天 API 端点使用 provider
src/services/models.ts：模型列表和模型拉取 API 端点使用 provider

为什么客户端模式之前能正常工作：
客户端模式不走服务器 API，而是直接在浏览器中通过 createPayloadWithKeyVaults(provider) 从 store 读取配置（使用原始 provider），然后使用 ModelRuntime.initializeWithProvider(runtimeProvider, ...) 选择 SDK（使用 sdkType）。两个参数职责清晰分离，不存在路径混淆问题。

🧪 How to Test

• Tested locally
• Added/updated tests
• No tests needed

📸 Screenshots / Videos

Before	After
...	...

📝 Additional Information

Summary by Sourcery

Ensure server-side AI API requests use the correct provider identifier for routing and configuration lookup.

Bug Fixes:

• Fix server-side chat streaming endpoint to address custom AI providers being unable to use their configured APIs.
• Fix model listing and model pull endpoints so custom AI providers resolve against the correct provider configuration instead of the SDK type.

[vercel]

@hardy-one is attempting to deploy a commit to the LobeHub OSS Team on Vercel.

A member of the Team first needs to authorize it.

[sourcery-ai]

Reviewer's guide (collapsed on small PRs)

Reviewer's Guide

Updates client-side chat and model service API calls to use the original provider identifier in endpoint paths instead of the runtime SDK type, so that server-side lookups fetch the correct custom provider configuration.

Sequence diagram for server chat request using provider instead of sdkType

sequenceDiagram
    actor User
    participant ChatUI
    participant ChatService
    participant API_ENDPOINTS
    participant Server
    participant ProviderConfigStore
    participant ModelRuntime

    User->>ChatUI: submitMessage(provider, sdkType, message)
    ChatUI->>ChatService: sendChat(provider, sdkType, payload)

    ChatService->>API_ENDPOINTS: chat(provider)
    API_ENDPOINTS-->>ChatService: /api/chat/{provider}

    ChatService->>Server: POST /api/chat/{provider} payload
    Server->>ProviderConfigStore: getConfig(provider)
    ProviderConfigStore-->>Server: ProviderConfig(sdkType, credentials,...)

    Server->>ModelRuntime: initializeWithProvider(ProviderConfig)
    ModelRuntime-->>Server: runtimeInstance

    Server->>ModelRuntime: sendMessage(payload)
    ModelRuntime-->>Server: streamingResponse
    Server-->>ChatService: SSE stream
    ChatService-->>ChatUI: deliverStream
    ChatUI-->>User: renderAssistantMessages

Loading

Sequence diagram for model list and pull using provider identifier

sequenceDiagram
    actor User
    participant ModelsUI
    participant ModelsService
    participant API_ENDPOINTS
    participant Server
    participant ProviderConfigStore
    participant ModelRuntime

    User->>ModelsUI: viewModels(provider, sdkType)
    ModelsUI->>ModelsService: listModels(provider, runtimeProvider, headers)

    alt useAgentRuntime
        ModelsService->>ModelRuntime: models()
        ModelRuntime-->>ModelsService: modelsList
    else useServerAPI
        ModelsService->>API_ENDPOINTS: models(provider)
        API_ENDPOINTS-->>ModelsService: /api/models/{provider}
        ModelsService->>Server: GET /api/models/{provider}
        Server->>ProviderConfigStore: getConfig(provider)
        ProviderConfigStore-->>Server: ProviderConfig(sdkType,...)
        Server->>ModelRuntime: initializeWithProvider(ProviderConfig)
        ModelRuntime-->>Server: modelsList
        Server-->>ModelsService: modelsList
    end

    ModelsService-->>ModelsUI: modelsList
    ModelsUI-->>User: showModels

    User->>ModelsUI: pullModel(provider, model)
    ModelsUI->>ModelsService: pullModel(provider, runtimeProvider, model, headers)

    alt useAgentRuntime
        ModelsService->>ModelRuntime: pullModel(model)
        ModelRuntime-->>ModelsService: pullResult
    else useServerAPI
        ModelsService->>API_ENDPOINTS: modelPull(provider)
        API_ENDPOINTS-->>ModelsService: /api/models/{provider}/pull
        ModelsService->>Server: POST /api/models/{provider}/pull { model }
        Server->>ProviderConfigStore: getConfig(provider)
        ProviderConfigStore-->>Server: ProviderConfig(sdkType,...)
        Server->>ModelRuntime: initializeWithProvider(ProviderConfig)
        ModelRuntime-->>Server: runtimeInstance
        Server->>ModelRuntime: pullModel(model)
        ModelRuntime-->>Server: pullResult
        Server-->>ModelsService: pullResult
    end

    ModelsService-->>ModelsUI: pullResult
    ModelsUI-->>User: showPullStatus

Loading

Updated class diagram for chat and model services provider handling

classDiagram
    class ChatService {
      +sendChat(provider, sdkType, payload, headers)
      +buildAnimations(responseAnimation)
    }

    class ModelsService {
      +listModels(provider, runtimeProvider, headers)
      +pullModel(provider, runtimeProvider, model, headers, signal)
    }

    class API_ENDPOINTS {
      +chat(provider)
      +models(provider)
      +modelPull(provider)
    }

    class ProviderConfigStore {
      +getConfig(provider)
    }

    class ProviderConfig {
      +provider
      +sdkType
      +credentials
      +settings
    }

    class ModelRuntime {
      +initializeWithProvider(providerConfig)
      +models()
      +pullModel(model)
      +sendMessage(payload)
    }

    ChatService --> API_ENDPOINTS : uses
    ChatService --> ModelRuntime : consumes via server

    ModelsService --> API_ENDPOINTS : uses
    ModelsService --> ModelRuntime : uses

    ProviderConfigStore --> ProviderConfig : returns
    ModelRuntime ..> ProviderConfig : depends on

    API_ENDPOINTS ..> ProviderConfig : keyed by provider
    ChatService ..> ProviderConfigStore : via server lookup
    ModelsService ..> ProviderConfigStore : via server lookup

Loading

File-Level Changes

Change	Details	Files
Use original provider ID instead of sdkType/runtimeProvider when constructing server API endpoints for chat and model operations.	Change chat SSE request to call API_ENDPOINTS.chat with provider rather than sdkType so server uses provider ID for configuration lookup. Change model list fetch to call API_ENDPOINTS.models with provider instead of runtimeProvider to align with server expectations. Change model pull request to call API_ENDPOINTS.modelPull with provider instead of runtimeProvider so custom providers hit their own endpoints.	src/services/chat/index.ts src/services/models.ts

Assessment against linked issues

Issue	Objective	Addressed	Explanation
#11271	Ensure that custom/self-hosted AI providers configured with an OpenAI-compatible format can correctly fetch the model list from the server API instead of failing or showing an incorrect default list.	✅
#11271	Ensure that chat/completion requests for custom AI providers (e.g., OpenAI-format, newapi with upstream octopus/gemini) are routed using the custom provider identifier rather than the sdkType, so they hit the correct backend configuration.	✅
#11296	Ensure that when using backend/server requests, chat and related AI requests are routed to the configured custom AI service provider instead of always going to the default OpenAI provider.	✅

Possibly linked issues

• AI服务商自建请求格式选择openAI貌似无效 #11271: The PR corrects endpoint provider resolution, fixing NEW API custom provider errors like the reported convert_request_failed.
• The custom AI service provider request failed #11296: PR changes API endpoint construction to use provider ID instead of sdkType, fixing custom backend provider routing regression.

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[gru-agent]

TestGru Assignment

Summary

Link	CommitId	Status	Reason
Detail	7e8ab3c	🚫 Skipped	No files need to be tested {"src/services/chat/index.ts":"Can not find valuable test target.\nChatService: Out of scope - The exported class ChatService only exposes methods as class properties, and the file exports an instance (chatService). Since the class contains only methods and no standalone exported functions, and its logic is highly coupled with multiple modules, external depende…","src/services/models.ts":"Can not find valuable test target.\nModelsService: Out of scope - The exported class ModelsService is primarily a service layer that handles fetching, downloading, and aborting model operations, including network requests and streaming responses. Its methods involve side effects, external dependencies, and process orchestration, mak…"}

History Assignment

Tip

You can @gru-agent and leave your feedback. TestGru will make adjustments based on your input

[sourcery-ai]

Hey - I've left some high level feedback:

• Given the subtle distinction between provider, runtimeProvider, and sdkType, consider tightening the types or renaming parameters in API_ENDPOINTS and call sites so it’s harder to accidentally pass the wrong identifier in the future.
• It may be helpful to add a small helper (e.g. getProviderEndpointId(provider) or similar) to centralize the mapping logic and avoid repeating the raw provider vs sdkType choice across multiple services.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- Given the subtle distinction between `provider`, `runtimeProvider`, and `sdkType`, consider tightening the types or renaming parameters in `API_ENDPOINTS` and call sites so it’s harder to accidentally pass the wrong identifier in the future.
- It may be helpful to add a small helper (e.g. `getProviderEndpointId(provider)` or similar) to centralize the mapping logic and avoid repeating the raw `provider` vs `sdkType` choice across multiple services.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[codecov]

Codecov Report

❌ Patch coverage is 66.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 76.40%. Comparing base (c0d165e) to head (5f172b7).
⚠️ Report is 4 commits behind head on next.

Additional details and impacted files

@@            Coverage Diff            @@
##             next   #11335     +/-   ##
=========================================
  Coverage   76.40%   76.40%             
=========================================
  Files        1134     1134             
  Lines       87549    87549             
  Branches    11796    10275   -1521     
=========================================
  Hits        66892    66892             
  Misses      20581    20581             
  Partials       76       76             

Flag	Coverage Δ
app	70.18% <66.66%> (ø)
database	94.08% <ø> (ø)
packages/agent-runtime	89.08% <ø> (ø)
packages/context-engine	83.19% <ø> (ø)
packages/conversation-flow	92.41% <ø> (ø)
packages/file-loaders	88.66% <ø> (ø)
packages/memory-user-memory	70.29% <ø> (ø)
packages/model-bank	100.00% <ø> (ø)
packages/model-runtime	86.88% <ø> (ø)
packages/prompts	76.00% <ø> (ø)
packages/python-interpreter	92.90% <ø> (ø)
packages/ssrf-safe-fetch	0.00% <ø> (ø)
packages/utils	92.88% <ø> (ø)
packages/web-crawler	95.62% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
Store	68.52% <ø> (ø)
Services	54.01% <66.66%> (ø)
Server	73.63% <ø> (ø)
Libs	41.53% <ø> (ø)
Utils	94.40% <ø> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Review	Updated (UTC)
lobehub	Ready	Preview, Comment	Jan 8, 2026 1:40pm

[lobehubbot]

❤️ Great PR @hardy-one ❤️

The growth of project is inseparable from user feedback and contribution, thanks for your contribution! If you are interesting with the lobehub developer community, please join our discord and then dm @arvinxx or @canisminor1990. They will invite you to our private developer channel. We are talking about the lobe-chat development or sharing ai newsletter around the world.

[lobehubbot]

🎉 This PR is included in version 2.0.0-next.244 🎉

The release is available on:

• npm package (@next dist-tag)
• GitHub release

Your semantic-release bot 📦🚀