Bin-Huang/camoufox-cli

Anti-detect browser automation CLI & Skills for AI agents — Camoufox-powered fingerprint spoofing, no bot-detectable Playwright leaks

What it solves

camoufox-cli provides a command-line interface for an anti-detect browser, allowing AI agents to browse the web without being detected as bots. It solves the problem of bot detection by spoofing hardware and software fingerprints (such as canvas, WebGL, and audio) at a C++ level, making the browser appear as a genuine human user.

How it works

The tool operates as a CLI that communicates via a Unix socket to a Python daemon. This daemon manages a Camoufox (Firefox-based) browser instance via Playwright. It allows agents to interact with pages using a simplified accessibility-tree snapshot system where elements are assigned reference IDs (e.g., @e1), which can then be used for clicking, filling forms, and extracting text.

Who it’s for

AI agents (such as Claude Code, Cursor, and OpenClaw) and developers who need to automate browser tasks while avoiding fingerprinting and bot detection systems.

Highlights

  • C++-level Fingerprint Spoofing: Spoofs canvas, WebGL, audio, screen metrics, and fonts to bypass detection.
  • Persistent Identities: Saves fingerprints, OS, and locale to a profile path, ensuring the same identity is maintained across different sessions.
  • Agent-Friendly Interaction: Uses accessibility-tree snapshots with element targeting via @ref for easy navigation and interaction.
  • Session and Tab Isolation: Supports multiple named sessions (separate browser processes) and named tabs (shared browser process) for parallel tasks.
  • Cookie Management: Ability to import and export cookies as JSON for session persistence.
  • Proxy and GeoIP Support: Integrated proxy support with automatic GeoIP spoofing based on the proxy IP.

Related

  • Project
  • Project
  • Project
  • Project
  • Project