hCaptcha Solver

Automated hCaptcha solving using AI vision models. Works with Discord and other sites that use hCaptcha.

What it does

This tool automatically solves hCaptcha challenges by:

• Using Multibot API for solving challenges
• Falling back to Groq's vision models when Multibot is unavailable
• Generating realistic mouse movement patterns
• Handling different challenge types (image selection, drag-drop, etc.)

Setup

Requirements

pip install tls-client camoufox colorama requests httpx asyncio jwt

You'll also need:

• A Groq API key (get one at console.groq.com)
• A Multibot API key (required for mouse movement generation)

Configuration

Edit config.json:

{
  "GROQ_API_KEY": "your-groq-api-key-here",
  "MULTIBOT_API_KEY": "your-multibot-key-here",
  "models": {
    "vision_model": "meta-llama/llama-4-scout-17b-16e-instruct",
    "fallback_model": "llama-3.1-8b-instant"
  },
  "debug": false
}

The Multibot API key is required for mouse movement generation. Without it, the solver won't work.

Usage

Basic example

from solver import HCaptchaSolver

solver = HCaptchaSolver(
    sitekey="a9b5fb07-92ff-493f-86fe-352a2803b3df",
    host="discord.com",
    rqdata=None,  # Optional: request data if required by the site
    real_time_mode=True
)

result = await solver.solve_captcha()

if result.get('success'):
    token = result.get('token')
    print(f"Got token: {token}")
else:
    print(f"Failed: {result.get('error')}")

With proxy

solver = HCaptchaSolver(
    sitekey="your-sitekey",
    host="example.com",
    rqdata=None,  # Optional: request data if required by the site
    proxy="username:password@ip:port"
)

Running the example

python solver.py

The main function in solver.py has a Discord example hardcoded. Just run it and it'll solve a challenge.

How it works

1. Gets site config - Fetches the hCaptcha configuration for the target site
2. Fetches challenge - Retrieves the actual challenge (images, questions, etc.)
3. Solves with AI - Uses Groq vision models or Multibot API to analyze and solve
4. Generates motion data - Creates realistic mouse movements for the submission
5. Submits solution - Sends the answer back to hCaptcha

The solver handles multiple challenge types:

• image_label_binary - Click all images that match
• image_label_area_select - Click on specific areas
• image_drag_drop - Drag objects to correct positions

Features

• Multiple AI backends - Groq for vision, Multibot for solving and mouse movements
• Realistic mouse movements - Generates human-like cursor paths
• Automatic retries - Handles failures and retries up to 15 times
• Debug logging - Clean, colorized logs when debug mode is enabled
• Proxy support - Works with HTTP proxies

Notes

• Success rate depends on your API keys and the challenge difficulty
• The solver uses a headless browser (Camoufox) for HSW token generation
• Multibot API key is required for mouse movement generation

Troubleshooting

"MULTIBOT_API_KEY not configured"

• You need a Multibot API key for mouse movement generation. Get one at multibot.in

Low success rate

• Try enabling Multibot API if you're only using Groq
• Some challenges are harder than others - the solver will retry automatically

HSW token generation fails

• Make sure you have Camoufox installed
• Check your proxy settings if using one

License

MIT License

Use at your own risk. This is for educational purposes.