Installation

Quick Start

The easiest way to get started is using the standalone server, which includes both the MCP endpoints and the web interface in a single process.

npx (Recommended)
Docker

If you have Node.js 22+ installed, run the server directly with a single command:

npx @arabold/docs-mcp-server@latest

This runs the server on port 6280 by default. Open http://localhost:6280 to access the web interface.

Use nvm use 22 and run npm rebuild if you recently changed Node versions.

With OpenAI Embeddings (Recommended):

OPENAI_API_KEY="sk-proj-..." npx @arabold/docs-mcp-server@latest

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

Running via Docker ensures you have all dependencies without polluting your host system:

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

The server writes its configuration to /config/docs-mcp-server/config.yaml. Mounting the /config volume ensures your settings persist across restarts.With OpenAI Embeddings:

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

Installation Methods

Embedded Server

Run the MCP server directly embedded in your AI assistant without a separate process or web interface. This provides MCP integration only. Add this to your MCP settings (VS Code, Claude Desktop, etc.):

{
  "mcpServers": {
    "docs-mcp-server": {
      "command": "npx",
      "args": ["@arabold/docs-mcp-server@latest"],
      "disabled": false,
      "autoApprove": []
    }
  }
}

When running in embedded mode, you lose access to the Web Interface unless you launch it separately.

Docker Compose (Scaling)

For production deployments or when you need to scale processing, use Docker Compose to run separate services.

Clone the repository

git clone https://github.com/arabold/docs-mcp-server.git
cd docs-mcp-server

Set environment variables

export OPENAI_API_KEY="your-key-here"

Start all services

docker compose up -d

Service Architecture:

Worker (port 8080): Handles documentation processing jobs
MCP Server (port 6280): Provides /sse endpoint for AI tools
Web Interface (port 6281): Browser-based management interface

Next Steps

Configuration

Configure the server with environment variables and config files

Connect Clients

Connect Claude, VS Code, Cursor, and other MCP clients

Getting Started

Setup

Guides

Architecture

Infrastructure

Quick Start

Installation Methods

Embedded Server

Docker Compose (Scaling)

Next Steps

Configuration

Connect Clients

Build docs developers (and LLMs) love

Getting Started

Setup

Guides

Architecture

Infrastructure

​Quick Start

​Installation Methods

​Embedded Server

​Docker Compose (Scaling)

​Next Steps

Configuration

Connect Clients

Build docs developers (and LLMs) love

Quick Start

Installation Methods

Embedded Server

Docker Compose (Scaling)

Next Steps