mcp
Verified Safeby jpotter80
Overview
A framework for building and deploying self-contained, searchable Model Context Protocol (MCP) servers from technical documentation with hybrid vector + keyword search for AI agents.
Installation
cd servers/mojo-manual-mcp && pixi run serveEnvironment Variables
- MAX_SERVER_URL
- EMBED_MODEL_NAME
- AUTO_START_MAX
- MOJO_DB_PATH
- MOJO_TABLE_NAME
- EMBED_CACHE_SIZE
- MAX_BINARY
- PYTHONPATH
Security Notes
The server's design inherently involves running external processes (`max serve` via `subprocess.Popen`) for embedding generation and syncing documentation from arbitrary Git repositories (via `sync_documentation.sh`). While these are intended for local development and user-initiated actions, they introduce potential attack vectors if untrusted URLs or commands are provided by the user. DuckDB's persistent secrets are explicitly noted as stored in unencrypted binary format on disk, which is a known characteristic of DuckDB itself, not a vulnerability introduced by this project.
Similar Servers
DevDocs
Provides intelligent web crawling and documentation extraction, storing content in a Model Context Protocol server for LLM querying and accelerating developer research.
Context-Engine
A Retrieval-Augmented Generation (RAG) stack for codebases, enabling context-aware AI agents for developers and IDEs through unified code indexing, hybrid search, and local LLM integration.
doc-mcp-server
Provides LLM agents with real-time, structured access to documentation from various package managers and platforms to prevent hallucinations and ensure up-to-date information.
RTFD
RTFD acts as an MCP server to provide LLM coding agents with real-time documentation from various package repositories and platforms, aiming to prevent hallucinations and provide up-to-date context.