vk_captcha - AI VK captcha solver for 93.2% accuracy

Requirements

Python3.3 - python3.10

Python 3.10 is not supported yet because onnxruntime is not supporting python3.10

UPDATE: Python3.10 is supported

Installation

pip install vk_captcha
or
pip install https://github.com/imartemy1524/vk_captcha/raw/main/dist/vk_captcha-2.0.tar.gz
or
pip install git+https://github.com/imartemy1524/vk_captcha

Fast examples:

Look into VkHacker for examples of accounts bruteforce

using vk_api library:

from vk_captcha import vk_api_handler
vk = vk_api_handler.VkApiCaptcha("88005553535", "efwoewkofokw")  # this login will create captcha
vk_api_handler.Solver.logging = True  # enable logging
vk.auth() # get captcha error and automatically solve it

another way with vk_api:

from vk_captcha import VkCaptchaSolver
from vk_api import VkApi
solver = VkCaptchaSolver(logging=True)  # use logging=False on deploy
vk = VkApi(login='', password='', captcha_handler=solver.vk_api_captcha_handler)
vk.method("any.method.with.captcha.will.be.handled")

using vkbottle:

from vkbottle.bot import Bot # Or "from vkbottle.user import User"
from vk_captcha import VkCaptchaSolver

bot = Bot(token=...) # Or "bot = User(token=...)"
solver = VkCaptchaSolver()

bot.api.add_captcha_handler(solver.vkbottle_captcha_handler)

just solve captcha from url / bytes

from vk_captcha import VkCaptchaSolver
import random, requests

session = requests.Session()  
session.headers['User-Agent'] = 'Mozilla/5.0 (Windows NT 10.0; rv:91.0) Gecko/20100101 Firefox/91.0'

solver = VkCaptchaSolver(logging=True)  # use logging=False on deploy
sid = random.randint(122112, 10102012012012)
easy_captcha = False
url = f"https://api.vk.com/captcha.php?sid={sid}&s={int(easy_captcha)}"

answer, accuracy = solver.solve(
    url=url,
    minimum_accuracy=0.33,  # keep solving captcha while accuracy < 0.33
    repeat_count=14,  # if we solved captcha with less than minimum_accuracy, then retry repeat_count times
    session=session  # optional parameter. Useful if we want to use proxy or specific headers
)
# or
#answer, accuracy = solver.solve(bytes_data=session.get(url))
print(f"I solved captcha = {answer} with accuracy {accuracy:.4}")

async way:

from vk_captcha import VkCaptchaSolver
import random, asyncio
solver = VkCaptchaSolver(logging=False)  # use logging=False on deploy
async def captcha_solver():
    sid = random.randint(122112, 10102012012012)
    easy_captcha = False
    url = f"https://api.vk.com/captcha.php?sid={sid}&s={int(easy_captcha)}"
    answer, accuracy = await solver.solve_async(url=url, minimum_accuracy=0.4, repeat_count=10)
    print(f"Solved captcha = {answer} with accuracy {accuracy:.4}")
asyncio.set_event_loop_policy(asyncio.WindowsSelectorEventLoopPolicy()) # Only in windows
asyncio.run(captcha_solver())

Also, you can get some statistics of solving captcha:

from vk_captcha import VkCaptchaSolver
solver = VkCaptchaSolver()
...
# solve some captchas
...
time_for1captcha = solver.argv_solve_time
total_solved = solver.TOTAL_COUNT
fail_count = solver.FAIL_COUNT  # you need directly increase it after getting second captcha error

In theory, for other languages you can use command line solver ( NOT RECOMMENDED, it will always load model again):

python -m vk_captcha -url "https://api.vk.com/captcha.php?sid=2323832899382092" -minimum-accuracy 0.33 -repeat-count 13