ContextMCP
A self-hosted MCP server that indexes documentation from various sources and serves it to AI Agents with semantic search.
ContextMCP
Self-hosted MCP server for your documentation. Index your documentation from across the sources and serve it via the Model Context Protocol (MCP) and REST API.
Quick Start
# Scaffold a new project
npx contextmcp init my-docs-mcp
# Follow the prompts, then:
cd my-docs-mcp
npm install
# Configure your API keys
cp .env.example .env
# Edit .env with your PINECONE_API_KEY and OPENAI_API_KEY
# Configure your documentation sources
# Edit config.yaml
# Index your documentation
npm run reindex
# Edit the cloudflare-worker
# Deploy the MCP server
cd cloudflare-worker
npm install
npm run deploy
What is ContextMCP?
ContextMCP creates a searchable knowledge base from your documentation that AI assistants can query via the Model Context Protocol (MCP).
Supported Content Types
| Parser | Content Types | Examples |
|---|---|---|
mdx | MDX/JSX documentation | Mintlify, Fumadocs, Docusaurus |
markdown | Plain Markdown files | READMEs, CHANGELOGs |
openapi | OpenAPI/Swagger specs | API reference docs |
How It Works
- Parse - Extract content from your docs, APIs, and READMEs
- Chunk - Split into semantic chunks optimized for search
- Embed - Generate embeddings using OpenAI
- Store - Upload to Pinecone vector database
- Search - Query via MCP from AI assistants
Repository Structure
contextmcp/
├── packages/
│ ├── cli/ # npx contextmcp (npm package)
│ ├── template/ # Project template (scaffolded to users)
│ └── website/ # contextmcp.ai documentation site
└── deployments/
└── dodopayments/ # Dodo Payments specific deployment
Packages
| Package | Description | Published |
|---|---|---|
packages/cli | CLI scaffolding tool | ✅ npm: contextmcp |
packages/template | Project template | (copied by CLI) |
packages/website | Documentation site | (deployed to Vercel) |
Development
Prerequisites
- Node.js 18+
Setup
# Install all dependencies
npm install
# Development
npm run dev:website # Run website locally
npm run dev:cli # Watch CLI for changes
# Build
npm run build:website # Build website
npm run build:cli # Build CLI
# Type checking
npm run typecheck # Check all packages
Documentation
Visit contextmcp.ai/docs for full documentation.
Contributing
We welcome contributions! Please see CONTRIBUTING.md for guidelines on how to contribute to this project.
License
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
Serveurs connexes
Compliance Auditor MCP
City hiring-compliance MCP server with regulation search and full audit risk scoring.
MCP Ripgrep Server
Provides local file search capabilities using the ripgrep (rg) command-line tool.
MCP Compass
Explore and discover Model Context Protocol servers using natural language queries.
Vistoya
Google for agentic fashion shopping/discovery. Indexed fashion brand e-coms. Semantic search.
Tripitaka MCP
Search and cite the full Pāli Canon (Tipiṭaka, ~444K segments) — Sutta, Vinaya, Abhidhamma at parity with SuttaCentral. Hybrid search, full-sutta fetch, translation comparison, Pāli word lookup. Free, non-commercial, offered as Dhamma Dāna.
Xiaohongshu XHS RedNote Social Media Data Assistant MCP
Read-only Xiaohongshu / XHS / RedNote MCP for note search, note details, paginated comments, creator profiles, and creator note lists.
Qdrant RAG MCP Server
A semantic search server for codebases using Qdrant, featuring intelligent GitHub issue and project management.
People Data Labs
Access person, company, school, location, job title, and skill data using the People Data Labs API.
MCP Knowledge Base
A knowledge base server that processes local documents (PDF, DOCX, TXT, HTML) and answers questions based on their content using similarity search.
SEC Filings and Earnings Call
The MCP server provides end-to-end workflows for SEC filings and earnings call transcripts—including ticker resolution, document retrieval, OCR, embedding, on-disk resource discovery, and semantic search—exposed via MCP and powered by the same olmOCR and embedding backends as the vLLM backends.