🤖 AutoCoder - Autonomous AI Coding Agent System

A powerful multi-agent AI system that autonomously generates complete software projects using state-of-the-art LLMs and intelligent agent orchestration.

✨ Latest Updates (August 2025)

🚀 API-Only Architecture

• Unified execution through centralized API server
• Automatic port detection and conflict resolution
• Embedded server management with zero configuration

🔑 Simplified Configuration

• No environment variables required! All API keys load from config.yaml
• Support for OpenAI, Anthropic, and Google Gemini
• Automatic provider detection and fallback

🧪 Comprehensive Testing

• Full CLI test coverage with mocked components
• Integration tests for all major features
• CI/CD pipeline with multi-version Python support

🎯 Key Features

• Multi-Agent System: Specialized agents for planning, development, testing, UI/UX, database, and DevOps
• LangGraph Orchestration: Intelligent workflow management with state persistence
• Auto Port Detection: Automatically finds available ports for API server
• Web Interface: Beautiful FastAPI-powered web UI for project management
• CLI Interface: Powerful command-line tools for developers
• MCP Support: Model Context Protocol integration for enhanced capabilities
• OpenAI Gateway: Compatible with OpenAI API for seamless integration

🚀 Quick Start

Installation

# Clone the repository
git clone https://github.com/eladrave/AutoCoder.git
cd AutoCoder

# Install dependencies
pip install -e .

Configuration

1. Copy the example config:

cp config.yaml.example config.yaml

2. Add your API keys to config.yaml:

api_keys:
  openai_api_key: "sk-your-key-here"
  anthropic_api_key: "sk-ant-your-key-here"
  google_api_key: "AIza-your-key-here"

No environment variables needed! The system loads everything from the config file.

Usage

Simple CLI Usage

# Create a simple application
./agent.sh "Create a todo list web app with React"

# Dry run mode (no files created)
./agent.sh "Build a REST API" --dry-run

# Verbose mode with logging
./agent.sh "Create a Python package" --verbose --save-logs

Advanced CLI Options

# Use specific port for API server
python main_api.py "Create app" --port 8080

# Use external API server
python main_api.py "Create app" --api-url http://remote-server:5001

# Keep server running for multiple tasks
python main_api.py "Create app" --keep-server

Web Interface

# Start the web interface
python web_interface/app.py

# Open browser to http://localhost:5001

📚 Documentation

• API Key Setup Guide - Configure API keys and providers
• Migration to API Guide - Transition to API-only architecture
• Cloud Deployment - Deploy to Google Cloud Run
• Monitoring Guide - System monitoring and observability

🧪 Testing

Run the comprehensive test suite:

# Run all tests
pytest tests/

# Run specific test categories
pytest tests/test_cli_api.py -v  # CLI tests
pytest tests/test_web_interface.py -v  # Web interface tests
pytest tests/test_agents.py -v  # Agent tests

# Run with coverage
pytest tests/ --cov=. --cov-report=html

🏗️ Architecture

┌─────────────────────────────────────────┐
│            Web Interface                 │
│         (FastAPI + React)                │
└────────────────┬────────────────────────┘
                 │
┌────────────────▼────────────────────────┐
│            API Server                    │
│     (Automatic Port Detection)           │
└────────────────┬────────────────────────┘
                 │
┌────────────────▼────────────────────────┐
│      Workflow Orchestrator               │
│         (LangGraph)                      │
└────────────────┬────────────────────────┘
                 │
     ┌───────────┴───────────┐
     │                       │
┌────▼──────┐       ┌────────▼───────┐
│  Agents   │       │    Services     │
├───────────┤       ├────────────────┤
│ • Planner │       │ • File Handler │
│ • Developer│      │ • Memory Store │
│ • Tester  │       │ • Config Loader│
│ • UI/UX   │       │ • Port Finder  │
│ • Database│       └────────────────┘
│ • DevOps  │
└───────────┘

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

📄 License

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

🙏 Acknowledgments

• Built with LangChain and LangGraph
• Powered by OpenAI, Anthropic, and Google AI models
• FastAPI for the robust web framework
• The amazing open-source community

📞 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: support@autocoder.ai

Built with ❤️ by the AutoCoder Team