Skip to main content

​
What is Camofox Browser?

Camofox Browser is a headless browser automation server built specifically for AI agents that need to browse the real web. It wraps the Camoufox engine (a Firefox fork with fingerprint spoofing at the C++ implementation level) in a REST API designed for token efficiency and reliable interaction.
Standing on the shoulders of Camoufox - The same engine behind askjo.ai’s web browsing.

​
Why Camofox exists

AI agents need to browse the real web. But traditional approaches fail:
  • Playwright gets blocked by modern bot detection
  • Headless Chrome gets fingerprinted immediately
  • Stealth plugins become the fingerprint themselves
Camoufox solves this by patching Firefox at the C++ implementation level:
  • navigator.hardwareConcurrency - spoofed in C++
  • WebGL renderers - patched before JS sees them
  • AudioContext - modified at the engine level
  • Screen geometry & WebRTC - all spoofed natively
No shims, no wrappers, no tells. JavaScript never knows it’s being watched.

​
Key features

C++ anti-detection

Bypasses Google, Cloudflare, and most bot detection systems by patching Firefox at the implementation level

Element refs

Stable e1, e2, e3 identifiers for reliable interaction - no brittle XPath or CSS selectors

Token-efficient

Accessibility snapshots are ~90% smaller than raw HTML, saving context window space

Runs on anything

Lazy browser launch + idle shutdown keeps memory at ~40MB when idle. Raspberry Pi, $5 VPS, shared Railway infra

Session isolation

Separate cookies and storage per user for multi-tenant agent systems

Cookie import

Inject Netscape-format cookie files for authenticated browsing without interactive login

Proxy + GeoIP

Route traffic through residential proxies with automatic locale/timezone matching

Search macros

Built-in shortcuts: @google_search, @youtube_search, @amazon_search, and 11 more

Structured logging

JSON log lines with request IDs for production observability

YouTube transcripts

Extract captions from any YouTube video via yt-dlp, no API key needed

Large page handling

Automatic snapshot truncation with offset-based pagination for content-heavy pages

Deploy anywhere

Docker, Fly.io, Railway, or standalone - runs wherever Node.js runs

​
Architecture overview

Camofox Browser uses a hierarchical session model: 
Browser Instance (Camoufox)
└── User Session (BrowserContext) - isolated cookies/storage
    ├── Tab Group (sessionKey: "conv1")
    │   ├── Tab (google.com)
    │   └── Tab (github.com)
    └── Tab Group (sessionKey: "conv2")
        └── Tab (amazon.com)
  • Browser instance: Single Camoufox process (launches on-demand, shuts down when idle)
  • User sessions: Isolated browser contexts per userId with separate cookies/storage
  • Tab groups: Tabs grouped by sessionKey for conversation/task isolation
  • Tabs: Individual browser tabs with stable element refs
Sessions auto-expire after 30 minutes of inactivity. The browser shuts down after 5 minutes with no active sessions, and relaunches on the next request.

​
How it works

Camofox exposes a REST API that AI agents use to control the browser:
  1. Create a tab → Returns tabId for subsequent operations
  2. Navigate → Go to URL or use search macro
  3. Get snapshot → Receive accessibility tree with element refs (e1, e2, etc.)
  4. Interact → Click/type using refs
  5. Repeat → Continue until task complete
The accessibility snapshot is the secret sauce: 
[heading] Example Domain
[paragraph] This domain is for use in examples.
[link e1] More information...
[button e2] Submit
Refs like e1 and e2 are stable identifiers that survive minor page changes - no brittle CSS selectors or XPath queries.

​
What’s different from other browser tools?

Playwright/Puppeteer use Chromium with stealth plugins. These get fingerprinted immediately by modern bot detection.Camofox patches Firefox at the C++ level before JavaScript ever runs. No shim layer to detect.

​
Get started

Quickstart

Get a working browser automation example in under 2 minutes

Installation

Install via npm, Docker, or as an OpenClaw plugin

API reference

Complete REST API documentation with all endpoints

GitHub

View source code and contribute on GitHub

​
License

Camofox Browser is MIT licensed. The underlying Camoufox engine is also open source - consider donating to its creator.