An intelligent multi-agent system built with Google Antigravity and Gemini 3 Pro (via Mistral API integration). This project demonstrates a hierarchical agent architecture where a "Lead Developer" orchestrates specialized sub-agents to handle design and coding tasks.
- Google Antigravity Framework: Utilizes the Antigravity agent structure for modular and scalable AI development.
- Gemini 3 Pro Intelligence: Powered by advanced LLMs (configured to use Mistral Large in this implementation) for high-quality reasoning and code generation.
- Specialized Sub-Agents:
- Designer: Equipped with browser capabilities to research design trends and create specifications.
- Coder: Equipped with code editor capabilities to generate production-ready HTML/CSS.
- Web Interface: A clean, chat-based UI for interacting with the agent system.
- Markdown Support: Responses are rendered with rich Markdown formatting.
- Node.js (v18 or higher)
- Mistral API Key (for AI model access)
-
Clone the repository:
git clone https://github.com/saketh8/Lead_Developer_Agent.git cd Lead_Developer_Agent -
Install dependencies:
npm install
-
Configure Environment: Create a
.envfile in the root directory and add your API key:MISTRAL_API_KEY=your_api_key_here
-
Start the Server:
node server.js
The server will start on
http://localhost:3000. -
Access the Agent: Open your browser and navigate to
http://localhost:3000. -
Interact:
- Design Queries: Ask about design trends (e.g., "Research 2025 web design trends"). The Designer agent will handle these.
- Coding Tasks: Ask for code (e.g., "Write HTML for a responsive navbar"). The Coder agent will handle these.
Lead_Developer_Agent/
├── agent.yaml # Agent configuration (roles, tools, models)
├── index.js # Agent entry point and orchestration logic
├── server.js # Express server for the web interface
├── public/ # Frontend assets
│ └── index.html # Chat interface
├── node_modules/ # Dependencies
├── .env # Environment variables (not committed)
└── README.md # Project documentation
Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License.