A stable, queue-based proxy for browser-use API with auto-retry and timeout protection.

• ✅ Queue management (one task at a time)
• ✅ Auto-retry (up to 3 attempts)
• ✅ Timeout protection (360 seconds)
• ✅ Beautiful web UI for monitoring
• ✅ Full result viewing with JSON output
• ✅ Task history and success rate tracking
• ✅ Screenshot support (when available)

This service is designed to be deployed on Render.com using the included render.yaml configuration.

1. Fork this repository
2. Go to Render Dashboard
3. Click "New +" → "Web Service"
4. Connect your forked repository
5. Render will auto-detect render.yaml and deploy

POST /api/submit
Content-Type: application/json

{
  "task": "Your task description",
  "llm_provider": "gemini",
  "max_steps": 20
}

GET /api/task/{task_id}

GET /api/status

GET /api/history

Access the web interface at the root URL to:

• Submit tasks via form
• View task history
• Monitor queue and processing status
• See full JSON results
• Track success rate

• REMOTE_API: The backend browser-use API URL (default: http://151.145.42.32:8000)
• LOCAL_PORT: Port to run on (default: 5001)

To prevent Render's free tier from spinning down, use a service like cron-job.org to ping the /api/status endpoint every 10 minutes.

MIT