A browser that forges its fingerprints in C++

What it does

Camofox-browser is a Node.js server that wraps the Camoufox engine—a Firefox fork with anti-detection patches applied at the C++ level—in a REST API built for AI agents. It is largely glue code: it exposes endpoints for navigation, interaction, and data extraction, returning accessibility snapshots that the project claims are roughly ninety percent smaller than raw HTML. Think of it as a headless-browser backend that aims to be a drop-in replacement for Puppeteer or Playwright stacks tired of being blocked.

The interesting bit

Instead of injecting JavaScript shims that bot detectors can fingerprint, Camoufox spoofs navigator.hardwareConcurrency, WebGL renderers, AudioContext, screen geometry, and WebRTC inside the browser implementation itself, before any page script runs. The server layer adds stable element references—literal e1, e2, e3 IDs—and accessibility snapshots to keep token usage low when feeding page state to an LLM.

Key highlights

• C++-level fingerprint spoofing via the Camoufox Firefox fork
• Accessibility snapshots claimed to be ~90% smaller than raw HTML
• Lazy browser launch and idle shutdown targeting ~40MB memory footprint when idle
• Built-in search macros for Google, YouTube, Amazon, Reddit, and others
• Anonymized crash/hang telemetry reported via GitHub Issues (opt-out)

Caveats

• Deployment has visible rough edges: the Dockerfile relies on bind mounts and pre-downloaded binaries, and the README carries strong warnings about line-ending issues on Windows and direct Docker builds
• Cookie import is disabled by default; the server rejects all cookie requests with a 403 unless an API key environment variable is set

Verdict

Worth a look if you are building an AI agent that needs to browse the modern web without being blocked by Cloudflare or bot detection, and you want a server-centric alternative to local Playwright. Skip it if you need a general-purpose scraping framework or prefer to manage browser binaries yourself.