Autonoma

A Multi-Agent AI Assistant System that lets you automate complex computational workflows via a simple chat interface over your local network. Built with a Python backend orchestrated by LangChain + LangGraph and a Next.js frontend, Autonoma empowers you to run tasks such as web scraping, scripting, file management, and full desktop automation—all within a secure, LAN-only environment.

🚀 Features

• Chat Interface: Interact via text or speech in English and Arabic.
• Multi-Agent Orchestration: Coordinator, Planner, Supervisor, Researcher, Coder, Browser, File Manager, Computer, Reporter.
• Real‐Time Transparency: Live activity stream, error reports, partial results.
• LAN‐Only Security: All traffic confined to your local network, sandboxed agents, full audit trails.
• Extensible: Easily add new agents or custom hooks.

📋 Prerequisites

• Python 3.9 or higher
• Node.js 16+ (for frontend build)
• pnpm (for agent-web dependencies)

🔧 Installation & Setup

1. Clone the repository

   git clone https://github.com/eslam-reda-div/Autonoma.git
   cd autonoma

2. Install Python dependencies

   pip install -r requirements.txt

3. Configure environment variables

   • Copy and edit backend .env

     cp .env.example .env

   • In .env, set:

• CHROME_INSTANCE_PATH → path to Chrome executable, e.g., "C:\Program Files\Google\Chrome\Application\chrome.exe"
• OPENWEATHERMAP_API_KEY → your OpenWeatherMap API key
• FIRECRAWL_API_KEY → your Firecrawl API key
• SERPER_API_KEY → your Serper API key
• TAVILY_API_KEY → your Tavily API key
• JINA_API_KEY → your Jina API key
• HYPERBROWSER_API_KEY → your HyperBrowser API key
• RIZA_API_KEY → your Riza API key

4. Configure credentials file

   cp credentials-example.json credentials.json

   • Open credentials.json and fill in any model-specific tokens or service credentials as needed the gpt-4o models use a github access token and to use the model from the github api if you want to use the official openai api then use the openai api key but change the MODEL_BASE_URL var in the env to the main openai base url, and the gemini models use a gemini api access token.

5. Build the frontend (agent-web)

   cd agent-web
   cp .env.example .env
   pnpm install
   pnpm build
   cd ..

▶️ Usage

1. Start the backend orchestrator

   python user.py

2. Scan the QR code displayed in your terminal or browser to connect your mobile/web client.
3. Interact with Autonoma via chat it to scrape data, run scripts, manage files, or automate desktop tasks.

🤝 Contributing

Contributions are welcome! Please open issues or pull requests to add new agents, improve documentation, or enhance functionality.