🐇 Rabbit

Rabbit is a modular, browser-controlling autonomous agent framework designed for intelligent web-based task execution. Leveraging LLMs and custom tools, Rabbit enables fully autonomous workflows such as research tasks, information extraction, and browser automation across complex multi-step processes.

---

🚀 Features

• 🔁 Agent loop execution (agent_task_loop.py)
• 🌐 Headless browser control with custom tools
• 🧠 LLM integration with memory + planning system
• 🔧 Extensible SDK (rabbit_sdk/) with modular components
• 🧪 Unit and workflow testing support
• 📊 Example workflows using real-world browser tasks

---

📼 Demo

---

📁 Project Structure

Rabbit/
├── agent_task_loop.py # Main agent loop runner
├── requirements.txt # Python dependencies
├── setup.py # Installation setup
├── test_agent.py # Agent test suite
├── rabbit_sdk/ # Core SDK for Rabbit agent
│   ├── __init__.py
│   ├── agent.py # RabbitAgent class definition
│   ├── browser_controller.py # Controls headless browser actions
│   ├── config.py # Environment + configuration loader
│   ├── llm_manager.py # LLM query & response handling
│   ├── memory_manager.py # Agent memory management
│   ├── planner.py # Task planning module
│   ├── tools/
│   │   ├── __init__.py
│   │   ├── browser_tools.py # Browser-specific tools
│   │   └── utility_tools.py # General utilities for agents
│   └── utils.py # Helper functions
└── examples/
    ├── __init__.py
    ├── simple_browser_task.py # Basic sentiment analysis example
    └── complex_workflow.py # Advanced multi-step automation example

---

🧪 Example: Sentiment Analysis of AI Safety

cd Rabbit/examples
python3 simple_browser_task.py

1. Open multiple URLs about AI and safety
2. Scrape relevant content
3. Run sentiment analysis
4. Summarize the key findings

---

🧪 Example: Comprehensive Crypto Analysis

cd Rabbit/examples
python3 complex_workflow.py

1. Open multiple URLs about crypto assets
2. Scrape relevant content
3. Run sentiment analysis
4. Summarize the key findings
5. Generate trading insights

---

⚙️ Setup & Installation

1. Clone the repository

git clone https://github.com/wchisasa/rabbit.git
cd rabbit

2. Install dependencies

pip install -r requirements.txt

3. Set environment variables

Create a .env file in the root with your API keys:

GEMINI_API_KEY=your_gemini_api_key_here

---

🧠 Powered By

• Google Gemini (or other LLMs)
• Playwright / Puppeteer (for browser automation)
• Open-source planning, memory, and tooling layers

---

🛠 Development & Testing

python test_agent.py

---

📌 TODO

• Add support for OpenAI + Claude
• Extend toolset for data transformation tasks
• Integrate with vector DB for persistent memory
• Web UI for visualizing agent reasoning

---

📄 License

MIT License. See LICENSE for details.