Simple program used to automate the solving of hCaptchas on sites such as discord.com
I created this program because I ahdnothing to do, I orignaly had it for personal use I then decided to impliment it into some programs of mine before it's solve rate began to decrease Once I saw this decrease I stopped working on it and now about 2 months later I've decided to add a few small updates With these updates the solve rate should now be much higher, but not very good overall
This works by using selenium to open a browser and go to an hcaptcha demo site where it opens the captcha The program then clicks on the images it classifies as being correct based off what the prompt is Due to hCaptcha's updates it is not working too well anymore, so feel free to contribute to the project
You can launch start.bat which opens command prompt and starts the main.py
If you would liek to change what website it solves captchas for, open main.py in any IDE or text editor
Please go down to line 35: driver.get('https://accounts.hcaptcha.com/demo')
Replace: https://accounts.hcaptcha.com/demo with any site of your choice
The site the solver should solve captchas on must instantly display a captcha box
For example, if the site requires you to type information first, like a signup form, the solver wil not work
If you want to do something like make it solve after the form is filled out, you will need to impliment the solver into your code
You will then need to fill out the form, then have the program call the function you put the solver code into
Alertanitvely, you can simply put the solver into a class and do something like: from HCapSolver import solver
You can retrain the capsolver by first opening images.json
You then need to add a new line, or edit an existing one and replace the two objects
The two objects should be objects that look very simialr, but are different, but don't get too specific
The objects should be simialr looking, but different because it helps teh model understand that a duck and a swan are different
Once you have the images.json fileld out you need to launch imagescraper.py
This program will automaticly scrape the web for images based off what you put in images.json
The program will put these images in .../content/train/object1 & .../content/train/object2
Once you have all your training data images, you can launch train.bat which will start the train.py file
The train.py file will start going through all the different images and classifying them into different groups and set an index to each image group
You will then see the class_indices.json file and when you open it you will see a list of every folder within .../content/train/ and a number
There will also be a model.keras file which is used by the program to classify the images and then decide which class each image on the captcha belongs to
Me, AG597
You can contact me on discord at ag597 if you need anything