I require a software tool to automatically solve hCaptcha challenges at scale. As you may know, hCaptcha is a popular captcha service used by many websites to protect forms from spam and abuse. However, it can also present a hurdle for legitimate automated workflows.
The tool should be able to visually interpret captcha images and videos, and programmatically select the appropriate responses such as street signs, storefronts, and traffic lights. It must solve challenges with a high degree of accuracy in order to avoid bot detections. Speed is also important, as it needs to clear several captchas per minute.
Advanced computer vision and deep learning techniques will likely be needed to achieve human-level performance. The front-end should provide a simple API or CLI interface to pass challenges and return solutions. Historical solve rates and any failures must be tracked. I'm hoping to integrate it with various scripts and scrapers to automate certain online tasks while avoiding any bot blocking.
The successful bidder will need relevant experience in CAPTCHA solving, OpenCV, Tensorflow, Python and similar technologies. Source code can be optionally provided but is not required. Please provide details of your approach and assurance that it will comply with websites' terms of use. I'm happy to discuss requirements further with the top submissions.
