Camofox Browser is a browser server that AI agents call over a REST API to browse sites that block automated traffic. It runs on Camoufox, a Firefox fork that spoofs fingerprint signals such as hardware concurrency, WebGL renderers, AudioContext, screen geometry, and WebRTC at the C++ level, before page JavaScript can read them.
Instead of raw HTML, the API returns accessibility snapshots with stable element refs (e1, e2, and so on) that an agent clicks or types into, and these snapshots are about 90% smaller than the page HTML. Search macros open common sites directly (Google, YouTube, Amazon, Reddit, Wikipedia, LinkedIn, and others), and the Reddit macros return JSON.
Each user gets an isolated browser context with its own cookies and storage, grouped into tabs per conversation. Sessions persist across restarts, expire after 30 minutes of inactivity, and the browser shuts down when idle, keeping memory around 40 MB so the server can share a small machine. It ships as an OpenClaw plugin with tools for tabs, snapshots, clicks, typing, scrolling, screenshots, and cookie import, or as a standalone server deployed with Docker, Fly.io, or Railway.
Features
- Fingerprint spoofing: Camoufox patches Firefox at the C++ level to get past Google, Cloudflare, and most bot detection
- Element refs: accessibility snapshots label interactive elements with stable refs for clicking and typing
- Search macros: 14 macros for sites including Google, YouTube, Amazon, Reddit, LinkedIn, TikTok, and Twitch
- Structured extract: maps a JSON Schema onto snapshot refs to return structured data from a page
- Session isolation and persistence: separate cookies and storage per user, restored after browser restarts
- Cookie import and VNC login: loads Netscape-format cookie files or lets a person log in through noVNC and reuses the stored state
- Proxy and GeoIP: routes traffic through single or rotating proxies with locale, timezone, and geolocation derived from the proxy IP
- YouTube transcripts: extracts captions from YouTube videos without an API key
- Downloads, images, and screenshots: captures browser downloads, lists page images, and attaches screenshots to snapshots
- Large pages: truncates long snapshots and pages through them by offset
- Session tracing: optional per-session Playwright traces with screenshots, DOM snapshots, network, and console output
- Crash telemetry: anonymized crash and hang reports filed as GitHub issues, with an opt-out
