Skip to content

v0.0.2

Latest
Latest

Choose a tag to compare

View all tags
@will-fawcett will-fawcett released this 01 Jun 16:19
v0.0.2
75ec69f
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Highlights

  • Pluggable LLM provider layer (gee_mcp.server.llm): Google (Gemini Developer / Vertex AI), Anthropic, and OpenAI behind a common BaseLLM interface.
  • New @register_llm decorator + from_env classmethod per provider — adding a provider is one decorated subclass.
  • Optional on-disk response cache (JSONFileCache) injected via ResponseCache protocol; default is NullCache.
  • Reasoning passthrough where supported (Anthropic adaptive thinking, OpenAI reasoning summaries).
  • New test module tests/test_server/test_llm.py (24 tests, no network).

Breaking changes

  • gee_mcp.server.genai module removed. Import from gee_mcp.server.llm instead (init_llm_client, BaseLLM, OpenAILLM, AnthropicLLM, GoogleLLM, …).
  • LLM_PROVIDER (google | anthropic | openai) and LLM_NAME env vars are now required for the analysis / code-generation / coderun tools. Previously these defaulted implicitly to Gemini.
  • Provider-specific credentials must match LLM_PROVIDER: OPENAI_API_KEY, ANTHROPIC_API_KEY, or one of GEMINI_API_KEY / GOOGLE_API_KEY / VERTEXAI_PROJECT.

See updated .env.example and README.md for configuration.

Assets 2
Loading