Automated Cloudflare account creation with Workers AI API token generation — bypasses Turnstile CAPTCHA and Cloudflare WAF using headless browser automation.

This tool automates the entire lifecycle of creating Cloudflare accounts with Workers AI access:

1. 📧 Generate temp email — via disposable mail API (any mailserver with compatible endpoint)
2. 🔐 Sign up Cloudflare account — fill form, solve Turnstile CAPTCHA, submit
3. 🔑 Create Account API Token — with Workers AI (Read + Edit) permissions
4. ✅ Validate token — verify against Workers AI REST API
5. 💾 Save to JSON/TXT — email, password, account_id, api_token, validation status
6. 📊 Live dashboard — optional Rich real-time worker progress/logs/statistics
7. 🧩 9Router export/add — export valid keys to 9Router-friendly TXT and bulk-add them locally

Output example:

{
  "email": "cf12345@yourdomain.com",
  "password": "Cf*Ab3xK9$mQ",
  "account_id": "a1b2c3d4e5f6789012345678abcdef01",
  "api_token": "cfut_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "token_valid": true,
  "workers_ai_models": 60,
  "token_name": "workers-ai-auto",
  "status": "full",
  "created_at": "2026-07-03T23:00:00+00:00",
  "proxy_used": "direct"
}

This project uses nodriver's built-in Cloudflare helper:

await page.verify_cf()

The helper drives the Cloudflare/Turnstile challenge from the browser session and avoids brittle image-template or OS-click logic. Keep the same authenticated page object after signup; opening a fresh tab/browser can lose dashboard session state.

Requires: xvfb-run to provide a virtual display server in VPS/headless environments. When running as root, browser startup uses sandbox=False.

• OS: Linux (Ubuntu 22.04+ recommended)
• Display: Xvfb (xvfb-run command)
• Browser: Google Chrome (stable channel)
• Python: 3.10+
• RAM: ≥512MB per browser instance
• Disk: ≥2GB (Chrome + dependencies)

# Clone the repo
git clone https://github.com/iAm-182/bluk-cf.git
cd bluk-cf

# Run setup script
chmod +x scripts/setup.sh
sudo ./scripts/setup.sh

# Edit config
cp config.example.json config.json
nano config.json

Edit config.json:

{
    "mail_api": "https://your-mail-api.example.com/api/new_address",
    "mail_domains": ["yourdomain.com", "anotherdomain.com"],
    "proxy": null,
    "headless": false,
    "max_accounts": 10,
    "delay_between_accounts": 300,
    "retry_attempts": 3,
    "token_name": "workers-ai-auto",
    "token_permissions": ["Workers AI"],
    "token_expiry": "no-expiration",
    "output_file": "results.json"
}

# Create 1 account
xvfb-run --auto-servernum python main.py

# Create 5 accounts with proxy
xvfb-run --auto-servernum python main.py -n 5 -p "http://user:pass@host:port"

# Create 10 accounts, custom output + 9Router TXT export
xvfb-run --auto-servernum python main.py -n 10 -o my_accounts.json --export-txt keys.txt

# Run with 2 concurrent workers and Rich live dashboard
xvfb-run --auto-servernum python main.py -n 10 --workers 2 -p "http://user:pass@host:port"

# Export existing valid results for 9Router
python scripts/export_9router_txt.py -i results.json -o keys.txt

# Bulk-add exported keys into local 9Router (localhost:20128)
python scripts/add_to_9router.py -i keys.txt

# Validate an existing token
python main.py --validate-only --token cfut_xxx --account-id abc123

# Batch run with proxy rotation
./scripts/batch_runner.sh 20 proxies.txt

python main.py [OPTIONS]

Options:
  -n, --accounts N          Number of accounts to create (default: 1)
  -c, --config FILE         Config file path (default: config.json)
  -p, --proxy URL           HTTP proxy URL
  -o, --output FILE         Output JSON file (default: results.json)
  -d, --delay SECS          Delay between accounts (default: 300)
  --headless                Run in headless mode
  --retry N                 Retry attempts per account (default: 3)
  -w, --workers N           Concurrent account workers (default: 1)
  --no-dashboard            Disable Rich live dashboard
  --export-txt FILE         Export valid keys to 9Router-friendly TXT
  --validate-only           Only validate an existing token
  --token TOKEN             Token to validate (with --validate-only)
  --account-id ID           Account ID for validation

Yes, but with caveats:

# 1. Prepare proxy list (residential, rotating)
# Format: one proxy per line
# http://user:pass@host:port

# 2. Use batch runner with proxy rotation
./scripts/batch_runner.sh 1000 proxies.txt

# 3. Or schedule via cron (recommended)
# Run 50 accounts every 6 hours
xvfb-run --auto-servernum python main.py -n 50 -p "http://user:pass@host:port" -d 600

┌─────────────────────────────────────────────┐
│           Scheduler (cron/systemd)          │
│  Runs every 6h, creates 50 accounts/run     │
└──────────────────┬──────────────────────────┘
                   │
        ┌──────────┼──────────┐
        ▼          ▼          ▼
    Proxy 1    Proxy 2    Proxy 3
        │          │          │
        ▼          ▼          ▼
    Browser    Browser    Browser
        │          │          │
        └──────────┼──────────┘
                   ▼
            results.json (append)

Key optimizations:

1. Sequential, not parallel — one browser at a time (memory efficient)
2. Proxy rotation — different IP per account
3. Scheduled runs — spread over hours to avoid rate limits
4. Append mode — results.json accumulates across runs
5. Retry logic — auto-retry on transient failures

cloudflare-auto-signup/
├── main.py                      # Entry point — orchestrator
├── config.example.json          # Config template (copy to config.json)
├── requirements.txt             # Python dependencies
├── README.md                    # This file
├── LICENSE                      # MIT License
├── .gitignore                   # Git ignore rules
├── src/
│   ├── __init__.py              # Package init
│   ├── email_generator.py       # Temp email API client
│   ├── turnstile_bypass.py      # OpenCV-based Turnstile solver
│   ├── signup_flow.py           # Signup automation (form + Turnstile)
│   ├── token_creator.py         # Account API Token creation
│   ├── token_validator.py       # Token validation via REST API
│   └── utils.py                 # Shared utilities
├── scripts/
│   ├── setup.sh                 # VPS setup (Chrome, Xvfb, deps)
│   └── batch_runner.sh          # Batch run with proxy rotation
├── docs/
│   ├── RATE_LIMITS.md           # Rate limit analysis & recovery times
│   ├── WAF_BYPASS.md            # WAF bypass techniques (detailed)
│   └── ARCHITECTURE.md          # Technical architecture diagram
└── tests/
    └── test_token_validator.py  # Validation tests

This tool is provided for educational and security research purposes only. Users are responsible for complying with Cloudflare's Terms of Service and all applicable laws. The authors are not responsible for any misuse.

• Auto-FreeCF — Original baseline concept and automation approach
• nodriver — Undetected Chrome automation
• Boterdrop-Solver — Camoufox CAPTCHA solver (cf_clearance)
• chatgpt-auto-signup — verify_cf() implementation reference
• OpenCV — Computer vision for template matching

This usually means:

1. Token was created but permissions weren't applied — re-run and check dashboard manually
2. Rate limit hit during token creation — account exists but token is incomplete
3. Token expired immediately — Cloudflare sometimes invalidates auto-created tokens

Workaround: Even if validation fails, the account_id + api_token are still saved in results.json. You can manually verify at https://dash.cloudflare.com/{account_id}/api-tokens.

# Check if Chrome is installed
google-chrome --version

# Check if Xvfb is installed
which xvfb-run

# Manual test — should open a browser window (or blank screen if no display)
xvfb-run --auto-servernum python -c "import nodriver as uc; import asyncio; asyncio.run(uc.start())"

# Update system
sudo apt update && sudo apt upgrade -y

# Install dependencies
sudo apt install -y xvfb google-chrome-stable python3.10 python3-pip git

# Clone
git clone https://github.com/iAm-182/bluk-cf.git
cd bluk-cf

# Python setup
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

You need a temporary email API that:

• Accepts POST /api/new_address with {"domain": "yourdomain.com"}
• Returns {"address": "user@yourdomain.com", "jwt": "..."}

Options:

• Self-hosted temp mail — Self-hosted disposable mail server (recommended)
• Mailinator API — Commercial, limited free tier
• Any disposable mail server — As long as it matches the API format

Configure in config.json:

{
    "mail_api": "https://your-mail-api.example.com/api/new_address",
    "mail_domains": ["yourdomain.com"]
}

cp config.example.json config.json
nano config.json  # Edit mail_api, mail_domains, proxy (if needed)

# Single account (recommended first run)
xvfb-run --auto-servernum python main.py

# Multiple accounts
xvfb-run --auto-servernum python main.py -n 5

# With proxy
xvfb-run --auto-servernum python main.py -n 5 -p "http://user:pass@host:port"

# Custom output file
xvfb-run --auto-servernum python main.py -n 10 -o my_accounts.json

# Export valid keys for 9Router while running
xvfb-run --auto-servernum python main.py -n 10 --export-txt keys.txt

# Add exported keys to local 9Router
python scripts/add_to_9router.py -i keys.txt

# View results
cat results.json | python -m json.tool

# Or use jq for filtering
cat results.json | jq '.[] | {email, account_id, api_token, status}'

Each account produces a token like:

{
  "account_id": "a1b2c3d4...",
  "api_token": "cfut_xxxxxxxxxxxxx"
}

Use it with Cloudflare Workers AI:

curl "https://api.cloudflare.com/client/v4/accounts/ACCOUNT_ID/ai/models/search" \
  -H "Authorization: Bearer cfut_XXXXXXXX"

MIT License — see LICENSE for details.