Build Your AI's Knowledge Base with Web Crawling and Document Management

Quick Start • Cursor Setup • Documentation

Archon is a Model Context Protocol (MCP) server that creates a centralized knowledge base for your AI coding assistants. Connect Cursor, Windsurf, or Claude Desktop to give your AI agents access to:

• Your documentation (crawled websites, uploaded PDFs/docs)
• Smart search capabilities with advanced RAG strategies
• Task management integrated with your knowledge base
• Real-time updates as you add new content

• Docker Desktop
• Supabase account (free tier works)
• OpenAI API key

# Clone the repository
git clone https://github.com/coleam00/archon.git
cd archon

# Create environment file
cp .env.example .env

Edit .env and add your Supabase credentials:

# Required
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_SERVICE_KEY=your-service-key-here

1. Set Up Database: In your Supabase project SQL Editor, run:

   -- Copy and paste the contents of migration/1_initial_setup.sql

2. Start Archon:

   docker-compose up --build -d

3. Configure API Key:

   • Open http://localhost:3737
   • Go to Settings → Add your OpenAI API key
   • Test by uploading a document or crawling a website

1. Add Project Management: In Supabase SQL Editor, run:

   -- Copy and paste the contents of migration/2_archon_projects.sql

2. Restart Python Server:

   docker-compose restart archon-pyserver

3. Enable Projects Feature:

   • Go to Settings in the web interface
   • Toggle "Enable Projects Feature" to ON
   • Access projects at http://localhost:3737/projects

1. Add MCP Client Features: In Supabase SQL Editor, run:

   -- Copy and paste the contents of migration/3_mcp_client_management.sql

2. Restart Services:

   docker-compose restart

3. Configure MCP Clients:

   • Access MCP Dashboard at http://localhost:3737/mcp
   • Add and manage MCP client connections

If you need to completely reset your database and start fresh:

Add this configuration to your Cursor settings:

File: ~/.cursor/mcp.json

{
  "mcpServers": {
    "archon": {
      "command": "docker",
      "args": [
        "exec", 
        "-i",
        "-e", "TRANSPORT=stdio",
        "-e", "HOST=localhost", 
        "-e", "PORT=8051",
        "archon-pyserver",
        "python", "src/mcp_server.py"
      ]
    }
  }
}

Complete documentation is available at: http://localhost:3838

Once everything is running:

1. Test Document Upload: Go to http://localhost:3737 → Knowledge Base → Upload a PDF
2. Test Web Crawling: Knowledge Base → "Crawl Website" → Enter a docs URL
3. Test Projects: Projects → Create a new project and add tasks
4. Test AI Integration: MCP Dashboard → Copy connection config for your AI client

• Smart Web Crawling: Automatically detects sitemaps, text files, or webpages
• Document Processing: Upload PDFs, Word docs, markdown, and text files
• AI Integration: Connect any MCP-compatible client (Cursor, Windsurf, etc.)
• Task Management: Organize projects and tasks with AI agent integration
• Real-time Updates: WebSocket-based live progress tracking

For development with hot reload:

# Backend (with auto-reload)
docker-compose up archon-pyserver --build

# Frontend (with hot reload) 
cd archon-ui-main && npm run dev

# Documentation (with hot reload)
cd docs && npm start

MIT License - see LICENSE file for details.

Build once, query everywhere
Transform your AI coding experience with Archon