README.md

Docs MCP Server

Maintained by Lab10 — helping companies drive AI adoption and increase their value.

Docs MCP Server solves the problem of AI hallucinations and outdated knowledge by providing an always-current documentation index for your AI coding assistant. It fetches official docs from websites, GitHub, npm, PyPI, and local files, allowing your AI to query the exact version you are using.

This project is a fork of arabold/docs-mcp-server by Andre Rabold, licensed under MIT. Lab10 extends it with additional features tailored to enterprise deployment needs.

Why Docs MCP Server?

• Up-to-Date Context: Fetches documentation directly from official sources on demand.
• Version-Specific: Queries target the exact library versions in your project.
• Reduces Hallucinations: Grounds LLMs in real documentation.
• Multiple Sources: Index websites, GitHub repositories, local folders, and zip archives.
• Rich File Support: Processes HTML, Markdown, PDF, Word (.docx), Excel, PowerPoint, and source code.
• Broad Compatibility: Works with any MCP-compatible client (Claude, Cline, Cursor, etc.).
• Production-Ready: Distributed deployment with auth, Railway config-as-code, and Docker Compose.

---

Deployment

Docker Compose

The recommended way to run a production instance with two services: a private worker (processes documentation) and a public MCP + API server (serves clients with JWT auth).

cp .env.production.example .env
# Edit .env with your real values (storage, auth, embeddings)
docker compose -f docker-compose.production.yml up -d

See .env.production.example for the full list of environment variables.

Railway

Config-as-code files are provided in the railway/ directory:

Service	Config	Networking
Worker	railway/worker.toml	Private (internal only)
MCP + API	railway/mcp.toml	Public (enable domain)

Setup:

1. Create a new Railway project linked to this repo.
2. Add two services, each pointing to its respective TOML config file.
3. Set environment variables per service (see .env.production.example for the Railway section).
4. The MCP service connects to the worker via Railway reference variables:

   WORKER_URL=http://${{worker.RAILWAY_PRIVATE_DOMAIN}}:${{worker.PORT}}/api
   

5. Volumes (dashboard only — not supported in TOML config): attach a volume to the worker service with mount path /data for persistent storage. Optionally mount /config for configuration persistence.

Docker (standalone)

For local development or single-machine deployments:

docker run --rm \
  -v docs-mcp-data:/data \
  -v docs-mcp-config:/config \
  -p 6280:6280 \
  ghcr.io/arabold/docs-mcp-server:latest \
  --protocol http --host 0.0.0.0 --port 6280

---

Connecting Clients

Add this to your MCP client settings (e.g., claude_desktop_config.json):

{
  "mcpServers": {
    "docs-mcp-server": {
      "type": "sse",
      "url": "http://localhost:6280/sse"
    }
  }
}

See Connecting Clients for VS Code (Cline, Roo) and other setup options.

---

Configuration

Embedding Model (Recommended)

Using an embedding model is optional but dramatically improves search quality by enabling semantic vector search.

OPENAI_API_KEY="sk-proj-..." docker compose -f docker-compose.production.yml up -d

See Embedding Models for configuring Ollama, Gemini, Azure, and others.

Storage Provider

By default, data is stored locally using SQLite. For team or cloud deployments, use Supabase (PostgreSQL):

DOCS_MCP_STORAGE_PROVIDER=supabase
DATABASE_URL="postgresql://user:pass@host:5432/db"
SUPABASE_SCHEMA=docs_prod  # optional, isolate tables into a custom schema

See Configuration for the full reference.

Authentication

Production deployments use OAuth2/OIDC JWT authentication to protect all public endpoints:

DOCS_MCP_AUTH_ENABLED=true
DOCS_MCP_AUTH_ISSUER_URL=https://your-auth-provider/.well-known/openid-configuration
DOCS_MCP_AUTH_AUDIENCE=urn:docs-mcp-server

See Authentication for details.

---

Documentation

• Installation: Setup guides for Docker and embedded mode.
• Connecting Clients: Claude, VS Code (Cline/Roo), and other MCP clients.
• Basic Usage: Using the Web UI, CLI, and scraping local files.
• Configuration: Full reference for config files and environment variables.
• Embedding Models: Configure OpenAI, Ollama, Gemini, and other providers.
• Deployment Modes: Standalone vs. Distributed (Docker Compose).
• Authentication: Securing your server with OAuth2/OIDC.
• Architecture: Deep dive into the system design.

---

Contributing

We welcome contributions! Please see CONTRIBUTING.md for development guidelines and setup instructions.

Acknowledgements

This project is based on docs-mcp-server originally created by Andre Rabold. We are grateful for his work in building the foundation of this tool.

License

This project is licensed under the MIT License. See LICENSE for details.