Adding Support for local OpenAI-compatible APIs #4
Conversation
- Add custom base URL support to OpenAI provider via OPENAI_BASE_URL env var - Add dedicated localopenai provider for better local API experience - Update documentation with local API setup instructions - Support popular local servers (LM Studio, Ollama, LocalAI, etc.) - Add .mcp-todos.json to gitignore
There was a problem hiding this comment.
Pull Request Overview
Adds support for local OpenAI-compatible APIs by extending the CLI’s LLM service and updating documentation.
- Introduce a new
localopenaiprovider inllm-service.tswith default endpoint and key fallback - Enhance the existing
openaiprovider to accept customOPENAI_BASE_URLenvironment variables - Extend
readme.mdwith detailed usage instructions and examples for local API servers
Reviewed Changes
Copilot reviewed 2 out of 3 changed files in this pull request and generated 2 comments.
| File | Description |
|---|---|
| source/services/llm-service.ts | Added localopenai provider and updated openai to support custom base URLs |
| readme.md | Documented usage for local OpenAI-compatible APIs with examples |
Comments suppressed due to low confidence (1)
source/services/llm-service.ts:45
- New
localopenaiprovider logic has been added without accompanying tests; consider adding unit tests to verify factory behavior under various environment and configuration scenarios.
localopenai: {
| factory: (key: string, cfg: LLMConfig) => | ||
| new ChatOpenAI({openAIApiKey: key, modelName: cfg.model}), | ||
| factory: (key: string, cfg: LLMConfig) => { | ||
| const baseURL = process.env['OPENAI_BASE_URL'] || process.env['OPENAI_API_BASE']; |
There was a problem hiding this comment.
[nitpick] The base URL normalization logic is duplicated between the openai and localopenai providers; consider extracting this into a shared helper to reduce duplication and improve maintainability.
Copilot uses AI. Check for mistakes.
| const baseURL = process.env['LOCAL_OPENAI_BASE_URL'] || 'http://localhost:1234/v1'; | ||
| const config: any = { | ||
| openAIApiKey: key || 'local-api-key', |
There was a problem hiding this comment.
[nitpick] Using a hardcoded fallback API key ('local-api-key') may lead to sending misleading credentials; consider allowing an empty key or requiring an explicit local key to avoid accidental misuse.
| const baseURL = process.env['LOCAL_OPENAI_BASE_URL'] || 'http://localhost:1234/v1'; | |
| const config: any = { | |
| openAIApiKey: key || 'local-api-key', | |
| if (!key) { | |
| throw new Error("API key for 'localopenai' provider must be explicitly provided."); | |
| } | |
| const baseURL = process.env['LOCAL_OPENAI_BASE_URL'] || 'http://localhost:1234/v1'; | |
| const config: any = { | |
| openAIApiKey: key, |
Copilot uses AI. Check for mistakes.
This pull request introduces support for local OpenAI-compatible APIs, enabling users to integrate local models and servers into the CLI tool. The changes include updates to the documentation, new configuration options for local APIs, and enhancements to the LLM service to handle these providers.
Documentation Updates:
readme.md: Added instructions for using local OpenAI-compatible APIs, including configuration options for setting base URLs and API keys for local servers like LM Studio, Ollama, LocalAI, and Text Generation WebUI.readme.md: Updated the list of supported models to include thelocalopenaiprovider for local APIs.Code Enhancements:
source/services/llm-service.ts: Added a newlocalopenaiprovider to thePROVIDERSobject. This includes logic for handling local API endpoints, default models, and API keys, with fallback values for local use cases. Updated theopenaiprovider to support custom base URLs via environment variables.