ContextMCP CLI

CLI tool to scaffold self-hosted MCP servers for your documentation. Create a searchable knowledge base from your docs that AI assistants can query via the Model Context Protocol (MCP) and REST API.

Installation

# Run directly with npx (recommended)
npx contextmcp init my-docs-mcp

# Or install globally
npm install -g contextmcp
contextmcp init my-docs-mcp

Usage

Initialize a New Project

npx contextmcp init [project-name]

The init command scaffolds a new ContextMCP project with everything you need to index your documentation and deploy an MCP server.

Interactive Mode

npx contextmcp init

Running without a project name will prompt you to enter one interactively.

Command Options

| Option | Description | | -------------- | ---------------------------- | | --no-install | Skip automatic npm install |

Example

# Create a new project
npx contextmcp init my-docs-mcp

# Navigate to project
cd my-docs-mcp

# Configure environment
cp .env.example .env
# Edit .env with your PINECONE_API_KEY and OPENAI_API_KEY

# Configure documentation sources
# Edit config.yaml to add your GitHub repos, docs, APIs

# Index your documentation
npm run reindex

# Deploy the MCP server
cd cloudflare-worker
npm install
npm run deploy

What Gets Scaffolded

When you run contextmcp init, you get a complete project structure:

my-docs-mcp/
├── src/
│   ├── parser/           # Document parsers (MDX, Markdown, OpenAPI)
│   ├── embeddings/       # OpenAI embedding generation
│   ├── sources/          # Source fetchers (GitHub, local, URL)
│   ├── config/           # Config schema and loader
│   └── types/            # TypeScript types
├── cloudflare-worker/    # MCP server deployment
├── scripts/              # Reindex and utility scripts
├── config.yaml           # Your configuration
├── config.example.yaml   # Example configuration
├── .env.example          # Environment template
└── package.json

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 |

Requirements

• Node.js 18+ is required
• Pinecone account for vector storage
• OpenAI API key for embeddings
• Cloudflare account for deployment (optional, for MCP server)

How It Works

1. Parse - Extract content from your docs, APIs, and READMEs
2. Chunk - Split into semantic chunks optimized for search
3. Embed - Generate embeddings using OpenAI
4. Store - Upload to Pinecone vector database
5. Search - Query via MCP from AI assistants

Documentation

For full documentation, visit contextmcp.ai/docs

Contributing

Contributions are welcome! Please see the main repository for contribution guidelines.

Related

• ContextMCP - Main repository
• Model Context Protocol - MCP specification

License

This project is licensed under the Apache-2.0 Licence. See the main LICENSE file for details.