Drop-in Playwright/Puppeteer replacement for Python and JavaScript. Same API, same code — just swap the import. Your browser now scores **0.9 on reCAPTCHA v3**, passes **Cloudflare Turnstile**, and clears **30 out of 30** stealth detection tests.
- **Config-level patches break** — `playwright-stealth`, `undetected-chromedriver`, and `puppeteer-extra` inject JavaScript or tweak flags. Every Chrome update breaks them. Antibot systems detect the patches themselves.
- **CloakBrowser patches Chromium source code** — fingerprints are modified at the C++ level, compiled into the binary. Detection sites see a real browser because it *is* a real browser.
- **One line to switch** — same Playwright API, no new abstractions, no CAPTCHA-solving services.
When using a proxy, antibot systems check that your browser's timezone and locale match the proxy's geographic location. CloakBrowser can auto-detect these from the proxy IP using an offline GeoIP database:
```bash
pip install cloakbrowser[geoip]# installs geoip2 + downloads ~70 MB database on first use
```
```python
# Timezone and locale auto-set from proxy's IP geolocation
> **Note:** The Playwright wrapper is recommended for sites with reCAPTCHA Enterprise. Puppeteer's CDP protocol leaks automation signals that reCAPTCHA Enterprise can detect, causing intermittent 403 errors. This is a known Puppeteer limitation, not specific to CloakBrowser. Use Playwright for best results.
Every launch automatically generates a **unique fingerprint**. A random seed (10000–99999) drives all seed-based patches — canvas, WebGL, audio, fonts, and client rects all produce consistent, correlated values derived from that single seed.
> **Tip: Use a fixed seed when revisiting the same site.** A random seed makes every session look like a different device — which can be suspicious when hitting the same site repeatedly from the same IP. For reCAPTCHA v3 Enterprise and similar scoring systems, a fixed seed produces a consistent fingerprint across sessions, making you look like a returning visitor:
| `--fingerprint-hardware-concurrency` | `8` | *(not set — uses real value)* | `navigator.hardwareConcurrency` |
| `--fingerprint-gpu-vendor` | `NVIDIA Corporation` | *(not set — native Apple GPU)* | WebGL `UNMASKED_VENDOR_WEBGL` |
| `--fingerprint-gpu-renderer` | `NVIDIA GeForce RTX 3070` | *(not set — native Metal renderer)* | WebGL `UNMASKED_RENDERER_WEBGL` |
> **Important:** `--fingerprint-platform` must always be set. The binary defaults to `windows` internally when this flag is missing, which causes GPU/UA mismatches on non-Windows systems. The wrapper handles this automatically.
> **Note:** All stealth tests were verified with the default fingerprint config above. Changing these flags may affect detection results — test your configuration before using in production.
### Examples
```python
# Default — unique fingerprint every launch
browser=launch()
# Pin a seed for a persistent identity
browser=launch(args=["--fingerprint=42069"])
# Full control — disable defaults, set everything yourself
**macOS (early access):** macOS builds are new — tested but not yet battle-tested at scale like Linux. If you hit any issues, [please open a GitHub issue](https://github.com/CloakHQ/CloakBrowser/issues).
**macOS first launch:** The binary is ad-hoc signed. On first run, macOS Gatekeeper will block it. Right-click the app → **Open** → click **Open** in the dialog. This is only needed once.
Some sites using advanced bot detection (e.g., DataDome, Cloudflare Turnstile) can detect headless mode even with our C++ patches. For these sites, run in **headed mode** with a virtual display:
```bash
# Install Xvfb (virtual framebuffer)
sudo apt install xvfb
# Start virtual display
Xvfb :99 -screen 0 1920x1080x24 &
exportDISPLAY=:99
```
```python
fromcloakbrowserimportlaunch
# Headed mode + residential proxy for maximum stealth
page.goto("https://heavily-protected-site.com")# passes DataDome, etc.
browser.close()
```
This runs a real headed browser rendered on a virtual display — no physical monitor needed. Combined with a residential proxy, this passes even the most aggressive detection services.
> **Tip:** Datacenter IPs are often flagged by IP reputation databases regardless of browser fingerprint. For sites with strict bot detection, a residential proxy makes the difference.
Avoid `page.wait_for_timeout()` — it sends CDP protocol commands that reCAPTCHA detects. Use native sleep instead:
```python
# Bad — sends CDP commands, reCAPTCHA detects this
page.wait_for_timeout(3000)
# Good — invisible to the browser
importtime
time.sleep(3)
```
```javascript
// Bad — sends CDP commands
awaitpage.waitForTimeout(3000);
// Good — invisible to the browser
awaitnewPromise(r=>setTimeout(r,3000));
```
Other tips for maximizing reCAPTCHA scores:
- **Use Playwright, not Puppeteer** — Puppeteer sends more CDP protocol traffic that reCAPTCHA detects ([details](#puppeteer))
- **Use residential proxies** — datacenter IPs are flagged by IP reputation, not browser fingerprint
- **Spend 15+ seconds on the page** before triggering reCAPTCHA — short visits score lower
- **Space out requests** — back-to-back `grecaptcha.execute()` calls from the same session get penalized. Wait 30+ seconds between pages with reCAPTCHA
- **Use a fixed fingerprint seed** (`--fingerprint=12345`) for consistent device identity across sessions
- **Use `page.type()` instead of `page.fill()`** for form filling — `fill()` sets values directly without keyboard events, which reCAPTCHA's behavioral analysis flags. `type()` with a delay simulates real keystrokes:
```python
page.type("#email", "user@example.com", delay=50)
```
- **Minimize `page.evaluate()` calls** before the reCAPTCHA check fires — each one sends CDP traffic
A: CloakBrowser is a browser. Using it is legal. What you do with it is your responsibility, just like with Chrome, Firefox, or any browser. We do not endorse violating website terms of service.
**Q: How is this different from Camoufox?**
A: Camoufox patched Firefox. We patch Chromium. Chromium means native Playwright support, larger ecosystem, and TLS fingerprints that match real Chrome. Also, Camoufox is no longer maintained (since March 2025).
**Q: Will detection sites eventually catch this?**
A: Possibly. Bot detection is an arms race. Source-level patches are harder to detect than config-level patches, but not impossible. We actively monitor and update when detection evolves.
**Q: Can I use my own proxy?**
A: Yes. Pass `proxy="http://user:pass@host:port"` to `launch()`.
