CamoFox

Anti-detection browser proxy for web scraping. Firefox-based (Camoufox fork) running on Fly.io.

Tech Stack

• Runtime: Node.js, Express, Playwright
• Browser: Camoufox v135.0.1-beta.24 (Firefox fork)
• Hosting: Fly.io (camofox app, iad region, 2GB RAM, auto-stop)
• Proxy: Thordata residential, US, 10GB balance (expires 31/03/2026)

URLs & Source

• Production: camofox.fly.dev
• Source: C:\Users\me\clawd\camofox-browser\server.js

API Endpoints

Method	Path	Description
GET	/health	Health check (503 if browser disconnected)
POST	/scrape	Scrape URL, returns HTML (asset blocking enabled)
POST	/render	JS-rendered page extraction (waitMs, waitFor, extractMode: text/markdown/html)
POST	/proxy-fetch	HTTP-level residential proxy requests (Thordata, rotating IPs)
POST	/cookies	Import cookies

Architecture

• Browser runs WITHOUT proxy (clean, fast)
• /proxy-fetch handles residential proxy at HTTP level via https-proxy-agent
• Asset blocking (images, fonts, CSS) on /scrape and /render to save bandwidth
• HTTP-level proxy design chosen because rotating residential proxy kills browser connections

Deployment

Deploy via flyctl:

C:\Users\me\.fly\bin\flyctl.exe deploy

Known Issues

• Facebook blocks datacenter IPs (use PinchTab with residential proxy instead)
• X/Twitter needs logged-in cookies for scraping