Skip to content
MCP server

Llmmcp

MCP server giving AI coding agents current OpenAI, Anthropic and Gemini API docs

Pricing
Open source
Platforms
CLI, API
Version
1.2.0
Listing updated
today

Similar MCP servers

About Llmmcp

Llmmcp is an open-source MCP server for developer tools. It is available on CLI and API.

The README says llmmcp is aimed at keeping coding assistants from relying on stale training data when they write code against LLM provider APIs. It exposes an MCP interface that agents such as Cursor, Claude Desktop or Windsurf can query for model names, tool-use syntax, context window sizes and rate limits from OpenAI, Anthropic and Google Gemini.

Under the hood, a weekly scraper collects official documentation, the text is chunked and embedded into a Pinecone index, and a Cloudflare Worker handles retrieval with Workers KV caching for frequent queries. A small CLI, published on npm, is what the MCP client actually launches.

It suits developers who build against these provider APIs and want their assistant to use current model identifiers and parameters. The project is MIT licensed and the author documents an option to run your own backend, though the deployment guide is marked as coming soon.

Tools this server provides (2)

Taken from the project documentation; the running server may differ.

search_docs
Search the latest official documentation for specific technical details.
list_providers
Get a dynamically updated list of available providers and their currently promoted models.

What it does

Search provider documentation
A search_docs tool retrieves passages from indexed official OpenAI, Anthropic and Gemini documentation.
Current model listing
A list_providers tool returns the providers covered and the models they currently promote.
Weekly documentation indexing
A scraper refreshes raw documentation text from official sources each week.
Cached query backend
A Cloudflare Worker serves retrievals and caches frequent results in Workers KV.
One-line client setup
Runs through npx with no local install, with configuration examples for Cursor and Claude Desktop.

Good for

  • Keep an AI coding assistant aware of the latest model names when writing LLM API calls
  • Look up tool-use syntax or context window limits for a specific provider model
  • Reduce deprecated library patterns in code generated for OpenAI, Anthropic or Gemini
  • Check rate limits before designing a batch processing job

Pricing

Open source

Free and MIT licensed; no paid plan mentioned.

Llmmcp FAQ

Which providers does llmmcp cover?

The README names OpenAI, Anthropic and Google Gemini as the providers whose official documentation is indexed.

How do I install it in Claude Desktop?

Add an mcpServers entry named llmmcp that runs npx with the arguments -y and llmmcp@latest in claude_desktop_config.json.

What license is it released under?

The README lists the MIT license.

Where is the vector data stored?

The README says chunks are embedded and stored in Pinecone, with query handling by a Cloudflare Worker.

Is Llmmcp free?

Yes. Llmmcp is open source, so you can use and self-host it for free.

What are the best alternatives to Llmmcp?

The closest alternatives to Llmmcp are awesome-mcp-servers, Proxima, awesome-mcp-servers, Api Test MCP and MCPfinder - MCP Server Discovery.

Where this comes from

  • Description: written from the project's README and its MCP Registry entry.

How AIPoweredBot checks servers