// Agent profile
Camofox Browser (jo-inc/camofox-browser) is an open-source anti-detection headless browser server purpose-built for AI agents. It wraps Camoufox, a Firefox fork that implements fingerprint spoofing at the C++ level — patching navigator.hardwareConcurrency, WebGL renderers, AudioContext, screen geometry, and WebRTC before JavaScript ever executes. This makes detection significantly harder than JavaScript-shim approaches used by most stealth browsers. The server exposes a REST API with stable element references (e1, e2, ...), accessibility snapshots ~90% smaller than raw HTML, session isolation, cookie injection, proxy routing with automatic GeoIP locale matching, and structured JSON logging. It runs on Node.js, installs with npm or Docker, and serves as a drop-in replacement for Puppeteer or Playwright when bot detection is a concern. MIT-licensed with 9.9K+ GitHub stars.
Camofox Browser is an anti-detection headless browser server designed for AI agents that need to browse the web without triggering bot detection. Unlike JavaScript-shim stealth approaches (which are increasingly detected), Camofox wraps Camoufox — a Firefox fork that implements fingerprint spoofing at the C++ level, deep inside the browser engine. Properties like navigator.hardwareConcurrency, WebGL renderer strings, AudioContext fingerprints, screen geometry, and WebRTC parameters are all patched before any page JavaScript executes, making fingerprint-based detection substantially harder.
The server exposes a REST API on port 9377, designed for AI agent consumption rather than traditional browser automation scripts. At nearly 10K GitHub stars, it has become a go-to choice for agent developers who need reliable web access in adversarial environments.
Engine-level anti-detection: The core differentiator. Rather than injecting JavaScript to override navigator properties (which detection scripts can catch by checking for overrides), Camoufox patches the browser engine itself. The fingerprint looks native because it is native — there’s no shim layer to detect.
Agent-optimized API: The REST interface returns stable element references (e1, e2, …) instead of fragile CSS selectors or XPath. Accessibility snapshots are approximately 90% smaller than raw HTML, reducing token consumption when feeding page state to LLMs.
Session management: Each session gets isolated cookies, storage, and a distinct fingerprint profile. Cookie injection via Netscape-format files enables pre-authenticated browsing without interactive login flows. Proxy routing automatically matches locale and timezone to the proxy’s GeoIP location for consistency.
Observability: Structured JSON logging and Playwright-compatible session traces (capturing network requests, DOM changes, and console output) make debugging autonomous browsing sessions practical.
AI agent web browsing: Any agent that needs to interact with websites protected by Cloudflare, Akamai, or similar bot detection systems. The REST API and compact snapshots are designed for LLM-driven navigation loops.
Research and data collection: Agents gathering information from sites that block headless browsers or rate-limit automated access. The session isolation and proxy support enable distributed collection across multiple identities.
Automated testing in adversarial environments: QA workflows where the system under test includes bot detection that interferes with standard Playwright/Puppeteer automation.
Camofox is a tool — its ethical use depends on the operator. Using it to bypass security measures on sites that prohibit automated access may violate terms of service. The ~300MB Camoufox browser download adds to deployment size. Performance overhead from the anti-detection patches is minimal but measurable versus vanilla Firefox. Residential proxies (not included) are recommended for highest-fidelity anti-detection.
AI agent developers who need reliable, undetected web browsing as part of autonomous workflows. Teams building web research agents, competitive intelligence tools, or any system where standard headless browsers get blocked. The MIT license and Docker deployment make it straightforward to integrate into existing agent infrastructure.
AI agents that autonomously conduct ML research experiments on single-GPU training, running a keep-or-revert loop overnight.
An AI-powered academic search engine that finds and synthesizes evidence-based answers from peer-reviewed scientific research.
An AI research assistant that helps researchers search, analyze, and synthesize findings from academic papers at scale.