CodeDox - AI-Powered Documentation Search & Code Extraction

Transform any documentation site into a searchable code database - CodeDox crawls documentation websites, intelligently extracts code snippets with context, and provides lightning-fast search via PostgreSQL full-text search and MCP (Model Context Protocol) integration for AI assistants.

📚 Documentation

For full documentation, installation guides, API reference, and more, visit:

https://chriswritescode-dev.github.io/codedox/

Quick Start

Docker Setup (Recommended)

# Clone the repository
git clone https://github.com/chriswritescode-dev/codedox.git
cd codedox

# Configure environment
cp .env.example .env
# Edit .env to add your CODE_LLM_API_KEY (optional for AI-enhanced extraction)

# Run the automated setup
./docker-setup.sh

# Access the web UI at http://localhost:5173

Manual Installation

See the full installation guide for detailed instructions.

Key Features

Intelligent Web Crawling: Depth-controlled crawling with URL pattern filtering and domain restrictions
Smart Code Extraction: Dual-mode extraction (AI-enhanced snippet descriptions or standalone)
Lightning-Fast Search: PostgreSQL full-text search with fuzzy matching
GitHub Repository Processing: Clone and extract documentation from GitHub repositories with full path support (e.g., /tree/main/docs)
MCP Integration: Native Model Context Protocol support for AI assistants
Modern Web Dashboard: React + TypeScript UI for visual management
Version Support: Track multiple versions of documentation
Real-time Monitoring: Live crawl progress and health monitoring
Upload Support: Upload documentation directly (Markdown, HTML, TXT) or from GitHub repositories

Demo - MCP Integration Example - OpenCode TUI

Screenshots

Dashboard

Markdown Search with Highlighting

Source Detail View

Contributing

See our Contributing Guide for details on how to contribute to CodeDox.

Author

Chris Scott - chriswritescode.dev

License

This project is licensed under the MIT License - see the LICENSE file for details.