hCaptcha Challenger
An open-source tool that solves hCaptcha challenges in an automated browser by pairing computer-vision models with a multimodal language model, without third-party anti-captcha services or Tampermonkey scripts.
hCaptcha Challenger automates hCaptcha's image challenges for browser agents such as Playwright, without relying on Tampermonkey scripts or paid third-party anti-captcha services. It implements its own interfaces so a locally run agent can face hCaptcha's own challenge system directly.
It handles several challenge types: binary image labeling with a ResNet ONNX classifier, area-selection challenges with YOLOv8 ONNX detection and segmentation models, multiple-choice labeling with a zero-shot ViT model, and drag-and-drop challenges through a spatial chain-of-thought approach. A CLIP-ViT model supports self-supervised challenge solving, and an agentic workflow layer coordinates these models together with a multimodal large language model.
It is written in Python and released under the GPL-3.0 license, with its own pipeline for collecting datasets and training and upgrading its detection models.