✨ Interactive Browser Use ✨

A web-based interface allowing users to direct, observe, and control an AI agent performing browser automation tasks in real-time with interactive step approval.

🎥 Demo Video

🚀 Take control of browser automation like never before! This project provides a hands-on web UI where you can collaborate with an AI agent, guiding its browsing actions step-by-step. Watch it navigate, interact with elements, and complete tasks, all under your supervision.

Key Features

• 👀 Real-time Observation: See exactly what the AI agent sees and does in the browser.
• ✅ Interactive Step Approval: Review and approve each action before the agent proceeds.
• ▶️ Direct Control: Guide the agent's direction and intervene when needed.
• 🌐 Web-Based Interface: Access and control the agent from your browser.
• 🧠 Powered by browser-use: Leverages the robust browser-use library (included in src/browser-use-src) for core agent capabilities.

Setup Instructions

Prerequisites

• Docker and Docker Compose

Installation

1. Clone the repository and navigate to the project directory:

   git clone https://github.com/Cofounder-Labs/interactive-browser-use
   cd interactive-browser-use

2. Copy the example environment file and add your API keys:

   cp .env.example .env

   Then edit .env and add your API keys:

   # Provide EITHER the OpenAI API Key OR the Azure configuration below
   
   # OpenAI API Key
   OPENAI_API_KEY=your_openai_api_key
   
   # --- OR ---
   
   # Azure OpenAI Configuration
   AZURE_ENDPOINT=https://your-resource-name.openai.azure.com/
   AZURE_OPENAI_API_KEY=your_azure_api_key
    
   

Running with Docker

1. Ensure Docker and Docker Compose are installed on your system.
2. Build and run the services using Docker Compose:

   docker compose up --build

   This command builds the images if they don't exist and starts the backend server and the Chrome instance in detached mode.
3. The application will be accessible at http://localhost:3000
4. To stop the services:

   docker compose down

Project Structure

interactive-browser-use/
├── .github/                # GitHub Actions workflows
├── frontend/               # Frontend application code (Next.js)
├── src/
│   ├── browser_agent/      # Backend FastAPI application and agent control logic
│   │   ├── web/            # FastAPI specific code (routes, models)
│   │   ├── utils/          # Utility functions
│   │   ├── agent.py        # Core agent interaction logic
│   │   └── cli.py          # Command-line interface (if applicable)
│   └── browser-use-src/    # Source code for the underlying browser-use library
├── .env.example            # Example environment variables
├── .gitignore
├── docker-compose.yml      # Docker Compose configuration
├── Dockerfile.backend      # Dockerfile for the backend service
├── Dockerfile.frontend     # Dockerfile for the frontend service
├── LICENSE
├── poetry.lock
├── pyproject.toml          # Python project configuration (Poetry)
├── README.md               # This file
└── run.py                  # Main entry point for running the backend locally

License

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

Acknowledgements

This project is built upon the excellent work of the browser-use team. We utilize their powerful browser-use library for the core browser automation capabilities. Many thanks to them for providing such a valuable open-source tool!

Contributors

• Ritanshu Dokania
• Re Solver

Citation

If you use Interactive Browser Use in your research or project, please cite:

@software{interactive_browser_use2025,
  author = {Dokania, Ritanshu and Solver, Re},
  title = {Interactive Browser Use: Make Browser Use Interactive},
  year = {2025},
  publisher = {GitHub},
  url = {https://github.com/Cofounder-Labs/interactive-browser-use}
}

Made with ❤️ in San Francisco