pinchtab

Active
GitHub Go MIT

Description

Pinchtab is a high-performance browser automation bridge for AI agents with multi-instance orchestration, stealth injection, and a real-time dashboard.

Key Features

  • Multi-instance orchestration - one control plane manages hundreds of isolated browser sessions
  • Stealth fingerprint injection - built-in anti-detection module bypasses popular anti-bot systems
  • Real-time dashboard - web UI shows session status, screenshots, and network traffic live
  • Standard agent interfaces - REST and WebSocket APIs let LangChain and other agents drive browsers directly
  • Containerized deployment - single docker-compose brings up the whole fleet
  • Session recording and replay - record browser sessions and replay them for regression testing

Use Cases

💡 Run hundreds of browser instances in parallel for AI agent price scraping across e-commerce sites
💡 Watch automation flows live in the dashboard and manually take over failed sessions
💡 Record real browser flows and replay them as end-to-end regression tests
💡 Use stealth mode for compliant red-team testing and research-grade data collection
💡 Expose the browser fleet as an internal SaaS for multiple agent teams

Strengths & Limitations

Strengths

  • Actively maintained, recent updates
  • High community interest (10.1k stars)
  • Permissive open-source license (MIT)
  • Responsive to issues, low backlog

Quick Start

# Launch the Pinchtab Docker fleet
git clone https://github.com/pinchtab/pinchtab.git
cd pinchtab
docker compose up -d

# Control plane listens on http://localhost:8080 by default
# Real-time dashboard at http://localhost:8080/dashboard

# Start a browser session via REST
curl -X POST http://localhost:8080/api/v1/sessions \
  -H 'Content-Type: application/json' \
  -d '{"profile": "default", "start_url": "https://example.com"}'

# Drive the browser over WebSocket from your agent
wscat -c ws://localhost:8080/api/v1/sessions/<id>/ws
# send: {"action":"navigate","url":"https://example.com/login"}

Related Projects

ScaleCUA

1.1k · Python
Stale C

Open-sourced computer use agents that can operate on cross-platform environments including Windows, macOS, Ubuntu, and Android. ICLR 2026 Oral paper project.

browserpythonagent +1
  • · Cross-platform GUI operation: supports Windows, macOS, Ubuntu, Android, and Web environments
  • · Large-scale training data: operation trajectories across 6 operating systems and 3 task domains
  • · State-of-the-art performance: MMBench-GUI L1-Hard 94.4%, OSWorld-G 60.6%, WebArena-Lite-v2 47.4%

PyWinAssistant

1.3k · Python
Stale C

The first open-source Artificial Narrow Intelligence generalist agent that fully operates GUIs using only natural language. Uses Visualization-of-Thought and Chain-of-Thought reasoning for spatial perception and HID simulation.

browseragentpython +2
  • · First open-source Computer-Using-Agent that operates GUIs through native Windows Accessibility APIs without OCR or pixel-level imaging
  • · Symbolic spatial mapping with hierarchical element tracking via OS-native parent/child relationships and coordinate systems
  • · Natural language to GUI automation: parse intent, map to UI elements, and generate adaptive workflows