Frequently Asked Questions (FAQ) about HexDocs MCP

Q: What is HexDocs MCP?

A: HexDocs MCP is a project that provides semantic search capabilities for Hex package documentation, designed specifically for AI applications. It consists of an Elixir binary for processing documentation and a TypeScript server implementing the Model Context Protocol (MCP) for searching.

Q: What is MCP (Model Context Protocol)?

A: MCP is an open protocol that standardizes how applications provide context to LLMs. The MCP server acts as a bridge, allowing AI models to access and interact with external data sources and tools.

Q: What are the requirements for using HexDocs MCP?

A: You need Ollama (with the nomic-embed-text model), Elixir 1.16+ and Erlang/OTP 26+, Mix, and Node.js 22 or later.

Q: How do I install HexDocs MCP?

A: You can install it by configuring your MCP client, using Smithery, or adding the hexdocs_mcp package to your Elixir project.

Q: How do I configure the embedding model?

A: You can configure the default embedding model using the HEXDOCS_MCP_DEFAULT_EMBEDDING_MODEL environment variable. Ensure Ollama is running and the specified model is pulled.

Q: How do I fetch documentation for a Hex package?

A: Use the command mix hex.docs.mcp fetch <package_name>. You can also specify the version and project path.

Q: How do I search for documentation?

A: Use the command mix hex.docs.mcp search <package_name> --query "<your_query>".

Q: Can I use HexDocs MCP offline?

A: Yes, documentation is stored locally, allowing offline access.

Q: How do I integrate HexDocs MCP with UBOS?

A: HexDocs MCP enhances UBOS by providing AI Agents with access to searchable Hex documentation, streamlining development and improving AI Agent training.

Q: Where is the documentation stored?

A: By default, data is stored in the ~/.hexdocs_mcp directory. You can change this using the HEXDOCS_MCP_PATH environment variable.

Q: What do I do if I get errors related to dependencies during installation?

A: Ensure all dependencies are available in the :dev environment. Update dependencies like Floki to be available in both :dev and :test environments.