panigale / selenium-recaptcha-solver

ReCAPTCHA solver for selenium (Using audio).

Selenium-recaptcha-solver

You can use this package to solve ReCAPTCHA challenges with selenium.

It supports single-step and multi-step audio solving for ReCAPTCHA V2 with >90% success rate.

Please use this package responsibly and for non-exploitative ends.

Notice: This project is no longer being maintained.

Requirements

Python 3.7+

Main dependencies:

• SpeechRecognition python package to transcribe speech
• Pydub for file conversions
• Requests for HTTP requests
• Selenium for web driver

If you're getting an error related to FFmpeg not being installed or in your PATH, get it here: https://ffmpeg.org/download.html

If the error persists, make sure FFmpeg is properly installed for your OS and in your PATH.

Installation

python -m pip install selenium-recaptcha-solver

Usage example with ReCAPTCHA demo site

from selenium_recaptcha_solver import RecaptchaSolver
from selenium.webdriver.common.by import By
from selenium import webdriver
from selenium.webdriver.chrome.options import Options


test_ua = 'Mozilla/5.0 (Windows NT 4.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/37.0.2049.0 Safari/537.36'

options = Options()

options.add_argument("--headless")  # Remove this if you want to see the browser (Headless makes the chromedriver not have a GUI)
options.add_argument("--window-size=1920,1080")

options.add_argument(f'--user-agent={test_ua}')

options.add_argument('--no-sandbox')
options.add_argument("--disable-extensions")

test_driver = webdriver.Chrome(options=options)

solver = RecaptchaSolver(driver=test_driver)

test_driver.get('https://www.google.com/recaptcha/api2/demo')

recaptcha_iframe = test_driver.find_element(By.XPATH, '//iframe[@title="reCAPTCHA"]')

solver.click_recaptcha_v2(iframe=recaptcha_iframe)

You can check a detailed use case in the tests folder of this project (Its execution is shown below in the demonstration chapter).

Demonstration

Avoiding detection

To decrease your chances of ReCAPTCHA detecting automated queries, try the following:

• Use a custom user-agent header (Make sure it's not a headless user-agent!)
• Use a hard-to-detect web driver
• Use proxies to avoid IP blacklisting

An example of a good user-agent: Mozilla/5.0 (Windows NT 4.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/37.0.2049.0 Safari/537.36

An example of a bad user-agent: are Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) HeadlessChrome/92.0.4512.0 Safari/537.36

Note the last part of the user-agent; the headless specification is usually there.

Questions

If the documentation hasn't covered something, or you have questions about how to use the package or how it works, please reach out.