Compare commits

...
3 Commits
Author SHA1 Message Date
CloakHQ 23ae521832 docs: add headed mode guide, JS roadmap, bump to 0.1.5
- Add Headed Mode section for DataDome/Cloudflare Turnstile sites
- Add Xvfb setup instructions + residential proxy tip
- Add JavaScript/Puppeteer support to roadmap
2026-02-23 18:13:07 +01:00
CloakHQ c2eb0ef21a feat: stealth hardening + test suite rewrite, bump to 0.1.4
- Remove --enable-automation via ignore_default_args (matches agent-browser
  Playwright patch — fixes connectionRTT, enables --fingerprint-* flags)
- Randomize fingerprint seed per launch (unique canvas/WebGL/audio per session)
- Change default GPU to RTX 3070 (higher market share = better blending)
- Rewrite stealth_test.py with JS evaluation for 6 detection sites
- Add --proxy flag to test script for Helper VPS routing
- All 6/6 tests passing (verified on VPS with v142 binary)
2026-02-23 07:58:50 +01:00
CloakHQ 2d7894ad68 docs: add Docker section to README, expand Dockerfile deps, bump to 0.1.3 2026-02-23 07:16:02 +01:00
8 changed files with 323 additions and 59 deletions
+4
View File
@@ -45,3 +45,7 @@ AGENTS.md
# Private docs (launch posts, strategy)
docs/
# Release scripts
publish.sh
.env
+5 -2
View File
@@ -1,11 +1,14 @@
FROM python:3.12-slim
# Playwright system deps
# Chromium system deps (matches fingerprint-chromium 142+ requirements)
RUN apt-get update && apt-get install -y --no-install-recommends \
libnss3 libnspr4 libatk1.0-0 libatk-bridge2.0-0 libcups2 \
libdbus-1-3 libdrm2 libxkbcommon0 libatspi2.0-0 libxcomposite1 \
libxdamage1 libxfixes3 libxrandr2 libgbm1 libpango-1.0-0 \
libcairo2 libasound2 libx11-xcb1 \
libcairo2 libasound2 libx11-xcb1 libfontconfig1 libx11-6 \
libxcb1 libxext6 libxshmfence1 \
libglib2.0-0 libgtk-3-0 libpangocairo-1.0-0 libcairo-gobject2 \
libgdk-pixbuf-2.0-0 libxss1 libxtst6 fonts-liberation \
&& rm -rf /var/lib/apt/lists/*
WORKDIR /app
+53 -15
View File
@@ -244,12 +244,64 @@ See the [`examples/`](examples/) directory:
| Linux x64 binary | ✅ Released |
| macOS arm64 (Apple Silicon) | 🔜 In progress |
| Chromium 145 build | 🔜 In progress |
| JavaScript/Puppeteer support (`cloakbrowser-js`) | 📋 Planned |
| Fingerprint rotation per session | 📋 Planned |
| Built-in proxy rotation | 📋 Planned |
| Windows support | 📋 Planned |
> ⭐ **Star this repo** to get notified when Chromium 145 and macOS builds drop.
## Docker
A ready-to-use [`Dockerfile`](Dockerfile) is included. It installs system deps, the package, and pre-downloads the stealth binary during build:
```bash
docker build -t cloakbrowser .
docker run --rm cloakbrowser python examples/basic.py
```
The key steps in the Dockerfile:
1. **System deps** — Chromium requires ~15 shared libraries (`libnss3`, `libgbm1`, etc.)
2. **`pip install .`** — installs CloakBrowser + Playwright
3. **`ensure_binary()`** — downloads the stealth Chromium binary at build time (~200MB), so containers start instantly
To extend with your own script, just add a `COPY` + `CMD`:
```dockerfile
FROM cloakbrowser
COPY your_script.py /app/
CMD ["python", "your_script.py"]
```
**Note:** If you run CloakBrowser inside a web server with uvloop (e.g., `uvicorn[standard]`), use `--loop asyncio` to avoid subprocess pipe hangs.
## Headed Mode (for aggressive bot detection)
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 &
export DISPLAY=:99
```
```python
from cloakbrowser import launch
# Headed mode + residential proxy for maximum stealth
browser = launch(headless=False, proxy="http://your-residential-proxy:port")
page = browser.new_page()
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.
## Troubleshooting
**Binary download fails / timeout**
@@ -264,20 +316,6 @@ You do NOT need `playwright install chromium`. CloakBrowser downloads its own bi
playwright install-deps chromium
```
**Missing system libraries on Linux (Docker)**
If you see errors about `libgbm`, `libnss3`, etc.:
```bash
apt-get install -y libgbm1 libnss3 libatk-bridge2.0-0 libxkbcommon0 libgtk-3-0
```
Or use `playwright install-deps chromium` which handles this automatically.
**Pre-download binary in Docker**
```python
# In your Dockerfile or entrypoint:
from cloakbrowser import ensure_binary
ensure_binary()
```
## FAQ
**Q: Is this legal?**
@@ -293,7 +331,7 @@ A: Possibly. Bot detection is an arms race. Source-level patches are harder to d
A: Yes. Pass `proxy="http://user:pass@host:port"` to `launch()`.
**Q: Can I use this with Docker?**
A: Yes. Use `ensure_binary()` in your Dockerfile to pre-download the binary during image build.
A: Yes. A ready-to-use Dockerfile is included — see the [Docker](#docker) section above.
## License
+2 -2
View File
@@ -12,7 +12,7 @@ Usage:
"""
from .browser import launch, launch_async, launch_context
from .config import CHROMIUM_VERSION, DEFAULT_STEALTH_ARGS
from .config import CHROMIUM_VERSION, get_default_stealth_args
from .download import binary_info, clear_cache, ensure_binary
from ._version import __version__
@@ -24,6 +24,6 @@ __all__ = [
"clear_cache",
"binary_info",
"CHROMIUM_VERSION",
"DEFAULT_STEALTH_ARGS",
"get_default_stealth_args",
"__version__",
]
+1 -1
View File
@@ -1 +1 @@
__version__ = "0.1.2"
__version__ = "0.1.5"
+4 -2
View File
@@ -17,7 +17,7 @@ from __future__ import annotations
import logging
from typing import Any
from .config import DEFAULT_STEALTH_ARGS
from .config import get_default_stealth_args
from .download import ensure_binary
logger = logging.getLogger("cloakbrowser")
@@ -63,6 +63,7 @@ def launch(
executable_path=binary_path,
headless=headless,
args=chrome_args,
ignore_default_args=["--enable-automation"],
**_build_proxy_kwargs(proxy),
**kwargs,
)
@@ -123,6 +124,7 @@ async def launch_async(
executable_path=binary_path,
headless=headless,
args=chrome_args,
ignore_default_args=["--enable-automation"],
**_build_proxy_kwargs(proxy),
**kwargs,
)
@@ -209,7 +211,7 @@ def _build_args(stealth_args: bool, extra_args: list[str] | None) -> list[str]:
"""Combine stealth args with user-provided args."""
result = []
if stealth_args:
result.extend(DEFAULT_STEALTH_ARGS)
result.extend(get_default_stealth_args())
if extra_args:
result.extend(extra_args)
return result
+13 -10
View File
@@ -4,6 +4,7 @@ from __future__ import annotations
import os
import platform
import random
from pathlib import Path
from ._version import __version__
@@ -17,16 +18,18 @@ CHROMIUM_VERSION = "142.0.7444.175"
# Default stealth arguments passed to the patched Chromium binary.
# These activate source-level fingerprint patches compiled into the binary.
# ---------------------------------------------------------------------------
DEFAULT_STEALTH_ARGS: list[str] = [
"--no-sandbox",
"--disable-blink-features=AutomationControlled",
# Fingerprint overrides (activate compiled C++ patches)
"--fingerprint=98765",
"--fingerprint-platform=windows",
"--fingerprint-hardware-concurrency=8",
"--fingerprint-gpu-vendor=NVIDIA Corporation",
"--fingerprint-gpu-renderer=NVIDIA GeForce RTX 4070",
]
def get_default_stealth_args() -> list[str]:
"""Build stealth args with a random fingerprint seed per launch."""
seed = random.randint(10000, 99999)
return [
"--no-sandbox",
"--disable-blink-features=AutomationControlled",
f"--fingerprint={seed}",
"--fingerprint-platform=windows",
"--fingerprint-hardware-concurrency=8",
"--fingerprint-gpu-vendor=NVIDIA Corporation",
"--fingerprint-gpu-renderer=NVIDIA GeForce RTX 3070",
]
# ---------------------------------------------------------------------------
# Platform detection
+241 -27
View File
@@ -1,57 +1,271 @@
"""Run stealth tests against major bot detection services.
Tests cloakbrowser against multiple detection sites and reports results.
Tests cloakbrowser against multiple detection sites, extracts pass/fail
verdicts via JS evaluation, and reports results with screenshots.
Usage:
python examples/stealth_test.py
python examples/stealth_test.py --headed # watch in real-time
python examples/stealth_test.py --no-screenshots
python examples/stealth_test.py --proxy http://10.50.96.5:8888
"""
import json
import sys
import time
from cloakbrowser import launch
HEADED = "--headed" in sys.argv
SCREENSHOTS = "--no-screenshots" not in sys.argv
PROXY = None
for i, arg in enumerate(sys.argv):
if arg == "--proxy" and i + 1 < len(sys.argv):
PROXY = sys.argv[i + 1]
def test_bot_sannysoft(page):
"""bot.sannysoft.com — classic bot detection checks."""
page.goto("https://bot.sannysoft.com", wait_until="networkidle", timeout=30000)
page.wait_for_timeout(3000)
results = page.evaluate("""() => {
const rows = document.querySelectorAll('table tr');
const data = {};
rows.forEach(r => {
const cells = r.querySelectorAll('td');
if (cells.length >= 2) {
const key = cells[0].innerText.trim();
const val = cells[1].innerText.trim();
const cls = cells[1].className || '';
data[key] = {value: val, passed: !cls.includes('failed')};
}
});
return data;
}""")
failed = [k for k, v in results.items() if not v["passed"]]
total = len(results)
passed = total - len(failed)
return {"passed": passed, "total": total, "failed": failed}
def test_bot_incolumitas(page):
"""bot.incolumitas.com — comprehensive 30+ check bot detection."""
page.goto("https://bot.incolumitas.com", wait_until="networkidle", timeout=30000)
page.wait_for_timeout(12000) # needs time to run all detection tests
# Site outputs JSON blocks in page text, not HTML tables
results = page.evaluate("""() => {
const text = document.body.innerText;
const okMatches = text.match(/"\\w+":\\s*"OK"/g) || [];
const failMatches = text.match(/"\\w+":\\s*"FAIL"/g) || [];
const failedTests = failMatches.map(m => m.match(/"(\\w+)"/)[1]);
return {
passed: okMatches.length,
failed: failMatches.length,
failedTests,
total: okMatches.length + failMatches.length
};
}""")
return results
def test_browserscan(page):
"""browserscan.net/bot-detection — WebDriver, UA, CDP, Navigator checks."""
page.goto("https://www.browserscan.net/bot-detection", wait_until="networkidle", timeout=30000)
page.wait_for_timeout(5000)
results = page.evaluate("""() => {
const items = document.querySelectorAll('[class*="result"], [class*="item"], [class*="check"]');
let normal = 0, abnormal = 0;
const text = document.body.innerText;
// Count "Normal" vs "Abnormal" verdicts
const normalMatches = text.match(/Normal/g);
const abnormalMatches = text.match(/Abnormal/g);
return {
normal: normalMatches ? normalMatches.length : 0,
abnormal: abnormalMatches ? abnormalMatches.length : 0,
pageText: text.substring(0, 500)
};
}""")
return results
def test_deviceandbrowserinfo(page):
"""deviceandbrowserinfo.com/are_you_a_bot — fingerprint + behavioral detection."""
page.goto("https://deviceandbrowserinfo.com/are_you_a_bot", wait_until="domcontentloaded", timeout=30000)
page.wait_for_timeout(8000)
results = page.evaluate("""() => {
const text = document.body.innerText;
// Site outputs JSON with "isBot": false and detail checks
const botMatch = text.match(/"isBot":\\s*(true|false)/);
const isBot = botMatch ? botMatch[1] === 'true' : null;
const checks = {};
const patterns = [
'isBot', 'hasBotUserAgent', 'hasWebdriverTrue',
'isHeadlessChrome', 'isAutomatedWithCDP', 'hasSuspiciousWeakSignals',
'isPlaywright', 'hasInconsistentChromeObject'
];
patterns.forEach(p => {
const match = text.match(new RegExp('"' + p + '":\\s*(true|false)'));
if (match) checks[p] = match[1] === 'true';
});
return {isBot, checks};
}""")
return results
def test_fingerprintjs(page):
"""demo.fingerprint.com/web-scraping — industry-standard bot detection."""
page.goto("https://demo.fingerprint.com/web-scraping", wait_until="networkidle", timeout=30000)
page.wait_for_timeout(5000)
# Click search to trigger bot detection — bots get blocked, humans see flights
try:
page.click("button:has-text('Search')", timeout=5000)
page.wait_for_timeout(5000)
except Exception:
pass
results = page.evaluate("""() => {
const text = document.body.innerText;
// Bots see error messages; humans see flight prices
const hasFlights = text.includes('Price per adult') || text.includes('$');
const isBlocked = text.includes('request was blocked') || text.includes('bot visit detected');
return {passed: hasFlights && !isBlocked, isBlocked, hasFlights};
}""")
return results
def test_recaptcha(page):
"""recaptcha-demo.appspot.com — Google's official reCAPTCHA v3 score."""
page.goto(
"https://recaptcha-demo.appspot.com/recaptcha-v3-request-scores.php",
wait_until="networkidle",
timeout=30000,
)
# Page auto-submits via grecaptcha.execute() — wait for backend response
page.wait_for_timeout(8000)
results = page.evaluate("""() => {
const text = document.body.innerText;
// Score appears in JSON response block: "score": 0.9
const scoreMatch = text.match(/"score":\\s*(\\d+\\.\\d+)/);
return {
score: scoreMatch ? parseFloat(scoreMatch[1]) : null,
pageText: text.substring(0, 500)
};
}""")
return results
TESTS = [
{
"name": "bot.sannysoft.com",
"url": "https://bot.sannysoft.com",
"runner": test_bot_sannysoft,
"verdict": lambda r: f"{r['passed']}/{r['total']} passed"
+ (f" (FAILED: {', '.join(r['failed'])})" if r["failed"] else " — ALL GREEN"),
"pass": lambda r: len(r["failed"]) == 0,
},
{
"name": "bot.incolumitas.com",
"url": "https://bot.incolumitas.com",
"check": "Bot detection analysis",
"runner": test_bot_incolumitas,
"verdict": lambda r: f"{r['passed']}/{r['total']} passed"
+ (f" (FAILED: {', '.join(r.get('failedTests', []))})" if r.get("failed", 0) > 0 else " — ALL GREEN"),
"pass": lambda r: r.get("failed", 0) <= 1, # fpscanner.WEBDRIVER false positive expected (all builds)
},
{
"name": "BrowserScan",
"url": "https://www.browserscan.net/bot-detection",
"check": "Bot detection status",
"runner": test_browserscan,
"verdict": lambda r: f"Normal: {r['normal']}, Abnormal: {r['abnormal']}",
"pass": lambda r: r.get("abnormal", 1) == 0,
},
{
"name": "deviceandbrowserinfo.com",
"url": "https://deviceandbrowserinfo.com/are_you_a_bot",
"check": "isBot flag",
"runner": test_deviceandbrowserinfo,
"verdict": lambda r: f"isBot: {r.get('isBot', 'unknown')}"
+ (f" checks: {json.dumps(r.get('checks', {}))}" if r.get("checks") else ""),
"pass": lambda r: not r.get("isBot", True),
},
{
"name": "FingerprintJS",
"url": "https://demo.fingerprint.com/web-scraping",
"check": "Bot detection result",
"runner": test_fingerprintjs,
"verdict": lambda r: "PASSED (flights shown)" if r.get("passed") else "BLOCKED" if r.get("isBlocked") else "NO FLIGHTS",
"pass": lambda r: r.get("passed", False),
},
{
"name": "reCAPTCHA v3 (Google)",
"url": "https://recaptcha-demo.appspot.com/recaptcha-v3-request-scores.php",
"runner": test_recaptcha,
"verdict": lambda r: f"Score: {r.get('score', 'N/A')}",
"pass": lambda r: (r.get("score") or 0) >= 0.7,
},
]
browser = launch(headless=True)
page = browser.new_page()
print("=" * 60)
print("CloakBrowser Stealth Test Suite")
print("=" * 60)
def main():
print("=" * 60)
print("CloakBrowser Stealth Test Suite")
print("=" * 60)
print(f"Mode: {'headed' if HEADED else 'headless'}")
print(f"Screenshots: {'on' if SCREENSHOTS else 'off'}")
print(f"Proxy: {PROXY or 'none'}")
print()
for test in TESTS:
print(f"\n--- {test['name']} ---")
print(f"URL: {test['url']}")
try:
page.goto(test["url"], wait_until="networkidle", timeout=30000)
page.wait_for_timeout(3000)
browser = launch(headless=not HEADED, proxy=PROXY)
page = browser.new_page()
# Screenshot each test
filename = f"stealth_test_{test['name'].replace('.', '_').replace(' ', '_')}.png"
page.screenshot(path=filename)
print(f"Screenshot: {filename}")
print(f"Title: {page.title()}")
except Exception as e:
print(f"Error: {e}")
results_summary = []
browser.close()
for test in TESTS:
name = test["name"]
print(f"--- {name} ---")
print(f"URL: {test['url']}")
print("\n" + "=" * 60)
print("Tests complete. Check screenshots for results.")
print("=" * 60)
try:
result = test["runner"](page)
passed = test["pass"](result)
verdict = test["verdict"](result)
status = "PASS" if passed else "FAIL"
results_summary.append((name, status, verdict))
print(f"Result: [{status}] {verdict}")
if SCREENSHOTS:
filename = f"stealth_test_{name.replace('.', '_').replace(' ', '_').replace('/', '_')}.png"
page.screenshot(path=filename)
print(f"Screenshot: {filename}")
except Exception as e:
results_summary.append((name, "ERROR", str(e)))
print(f"Error: {e}")
print()
browser.close()
# Summary table
print("=" * 60)
print("RESULTS SUMMARY")
print("=" * 60)
for name, status, verdict in results_summary:
icon = {"PASS": "+", "FAIL": "!", "ERROR": "x"}[status]
print(f" [{icon}] {name}: {verdict}")
passed_count = sum(1 for _, s, _ in results_summary if s == "PASS")
total = len(results_summary)
print(f"\n {passed_count}/{total} tests passed")
print("=" * 60)
return 0 if passed_count == total else 1
if __name__ == "__main__":
sys.exit(main())