Rust Docs MCP Server

What is Rust Docs MCP Server?

Rust Docs MCP Server is an MCP server that fetches documentation for a specified Rust crate, generates embeddings, and provides an MCP tool to answer questions about the crate based on the documentation context. It is designed for AI-powered coding assistants (such as Cursor, Cline, Roo Code) to access current, crate-specific API documentation, improving the accuracy of generated Rust code.

How to use Rust Docs MCP Server?

Install by downloading the pre-compiled binary for your OS from the GitHub Releases page (or build from source with the Rust Toolchain). Set the OPENAI_API_KEY environment variable. Run from the command line with a package ID specification (e.g., rustdocs_mcp_server serde@^1.0), optionally adding features with -F. The server communicates via stdio and exposes a tool query_rust_docs and a resource crate://<crate_name>. On first run for a new crate/version/feature set, it downloads docs and generates embeddings (then caches them). Subsequent launches use the cache.

Key features of Rust Docs MCP Server

• Targets a single Rust crate per server instance.
• Supports specifying required crate features for documentation generation.
• Uses OpenAI text-embedding-3-small for semantic search.
• Uses OpenAI gpt-4o-mini-2024-07-18 for LLM summarization.
• Caches documentation and embeddings in the XDG data directory.
• Runs as a standard MCP server over stdio.

Use cases of Rust Docs MCP Server

• AI assistant generates code using the latest API of a crate like serde or tokio.
• A developer configures multiple server instances for different crates (e.g., reqwest and async-stripe).
• An LLM queries crate-specific documentation before writing code, reducing outdated suggestions.
• Teams ensure generated Rust code aligns with current crate versions and feature flags.

FAQ from Rust Docs MCP Server

What is Rust Docs MCP Server?

It is an MCP server that provides an AI assistant with up-to-date, crate-specific documentation via semantic search and LLM summarization.

How can I use it with multiple crates?

Run multiple server instances concurrently, each targeting a different crate (using different package ID specifications). Configure them in your MCP client (e.g., Roo Code or Claude Desktop) with separate entries.

What are the prerequisites?

An OpenAI API key (set as OPENAI_API_KEY environment variable) and network access to download crate dependencies and interact with the OpenAI API.

How much does embedding generation cost?

Typically fractions of a US penny per crate. For example, a large crate like async-stripe with over 5000 documentation pages cost $0.18 USD during testing.

How does caching work?

Cached documentation and embeddings are stored in ~/.local/share/rustdocs-mcp-server/<crate_name>/<version>/<features_hash>/ using bincode. Different version requirements and feature sets are cached separately. If the cache is missing or corrupted, the server regenerates automatically.