I've decided to open source my old Recaptcha v2 solver. My latest version will be opened sourced this summer. I am hoping this project will serve as inspiration for others to build a solver from, as plugging in an existing model is not too difficult to do.

This project uses the AWS Rekognition API, which is a decent solution but isn't the best. If you are looking to build off this project, I would take a look at this repo for inspiration on building and training your own model. I would also reccomend using this repo as a base for your dataset or scraping images from Google Images using this project.

• Python 3
• Firefox
• Geckodriver (make sure it is installed in your PATH)
• AWS credentials (create a new IAM user with the AmazonRekognitionFullAccess role)

git clone https://github.com/natewong1313/recaptcha-fullauto.git

cd recaptcha-fullauto

Add your proxies in the proxies.txt file and add your AWS credentials in the .env file
If you don't want to use proxies, modify the src/main.py file as such

rcs = RecaptchaSolver("https://www.google.com/recaptcha/api2/demo", use_proxies = False)

docker build -t recaptcha-fullauto .

docker run recaptcha-fullauto

pip install -r requirements.txt

python src/main.py