ModelHub - AI Agent Supervisor & Guide

The intelligent AI agent supervisor that guides your coding assistants to success.

---

✨ What is ModelHub?

ModelHub is an intelligent AI agent supervisor that monitors and guides other AI coding assistants (like OpenCode, Cursor, Claude Code, etc.) to ensure high-quality, goal-oriented development.

Think of it as a senior engineer looking over the shoulder of your AI coding assistant, providing:

• 🎯 Goal-oriented supervision - Ensures every action aligns with your project objectives
• 🧠 Contextual memory - Remembers previous conversations and decisions
• ⚡ Dual modes - Choose between "Guidance Mode" (advice only) or "Execution Mode" (hands-on)
• 🔇 Error awareness - Intelligently detects and handles errors in guidance mode
• 🌐 Multi-terminal support - Works with iTerm2, tmux, and Windows Terminal

🚀 Quick Start

One-Command Installation

curl -fsSL https://agentcompass.ai/install | bash

Manual Installation

# Clone the repository
git clone https://github.com/yourusername/agentcompass.git
cd agentcompass

# Install dependencies
pip install -r requirements.txt

# Run ModelHub
python -m agentcompass.main

Configure Your API Key

Edit config.yaml:

# ModelHub Configuration
# Get your API key from https://platform.deepseek.com/

api_key: "your-api-key-here"
model: "deepseek-chat"
base_url: "https://api.deepseek.com"

# Or use environment variables:
# export AGENTCOMPASS_API_KEY="your-api-key"

🎯 Features

🎯 Goal-Oriented Supervision

Set your project goal once, and ModelHub ensures every AI action stays on track.

🧠 Contextual Memory

ModelHub remembers previous conversations, preventing repetitive advice and maintaining logical continuity.

⚡ Dual Operating Modes

Mode	Description	Best For
Guidance Mode 🧠	Provides feedback and suggestions only	Learning, code review
Execution Mode ⚡	Offers executable code and commands	Rapid prototyping

🌐 Universal Terminal Support

• iTerm2 (macOS)
• tmux (macOS/Linux/WSL)
• Windows Terminal (Windows)

🔧 Multi-Model Support

Built-in support for multiple LLM providers:

# config.yaml
providers:
  deepseek:
    enabled: true
    model: "deepseek-chat"
    
  openai:
    enabled: true
    model: "gpt-4"
    
  anthropic:
    enabled: true
    model: "claude-3-5-sonnet"

📖 Documentation

Full documentation is available at docs.agentcompass.ai

• Getting Started
• Configuration Guide
• API Reference
• Examples
• Contributing Guide

🛠️ Usage

Basic Usage

# Auto-detect terminal and start
python -m agentcompass.main

# Force specific terminal mode
python -m agentcompass.main --terminal tmux

Programmatic Usage

from agentcompass import DeepSeekClient, ModernInput

# Configure and run
client = DeepSeekClient(config_path="config.yaml")
input_ui = ModernInput()

goal, mode = input_ui.get_user_input()
print(f"Starting ModelHub in {mode} mode for: {goal}")

🤝 Contributing

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

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

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

🙏 Acknowledgments

• OpenCode - Inspiration for the project structure
• DeepSeek - For the excellent API
• All contributors and users of ModelHub

---

Made with ❤️ by the ModelHub Team

Website | GitHub | Documentation