bestcaptchasolver-python3 is a super easy to use bypass captcha python3 API wrapper for bestcaptchasolver.com captcha service
git clone https://github.com/bestcaptchasolver/bestcaptchasolver-python3
pip install -r requirements.txt
# make sure you've changed access_key, page_url, etc in main.py
python3 main.py
Simply require the module, set the auth details and start using the captcha service:
from bestcaptchasolverapi3.bestcaptchasolverapi import BestCaptchaSolverAPI
Set access_token for authentication
access_token = 'access_token_here'
# get your access token from https://bestcaptchasolver.com/account
bcs = BestCaptchaSolverAPI(ACCESS_TOKEN)
Once you've set your authentication details, you can start using the API
Get balance
balance = bcs.account_balance()
Submit image captcha
data = {}
data['image'] = 'captcha.jpg'
# optional parameters
data['is_case'] = if case sensitive set to True, default: False
data['is_phrase'] = if phrase, set to True, default: False
data['is_math'] = True if captcha is math, default: False
data['alphanumeric'] = 1 (digits only) or 2 (letters only), default: all characters
data['minlength'] = minimum length of captcha text, default: any
data['maxlength'] = maximum length of captcha text, default: any
bcs.submit_image_captcha(data)
The image submission works with both files and b64 encoded strings.
For setting the affiliate_id, set the affiliate_id
parameter
Submit recaptcha details
For recaptcha submission there are two things that are required.
- page_url
- site_key
- type (optional, defaults to 1 if not given)
1
- v22
- invisible3
- v34
- enterprise v25
- enterprise v3
- v3_action (optional)
- v3_min_score (optional)
- domain (optional) - i.e
www.google.com
orrecaptcha.net
- data_s (optional)
- cookie_input (optional)
- user_agent (optional)
- affiliate_id (optional)
- proxy (optional)
bcs.submit_recaptcha({'page_url': 'page_url_here', 'site_key': 'sitekey_here')
This method returns a captchaID. This ID will be used next, to retrieve the g-response, once workers have completed the captcha. This takes somewhere between 10-80 seconds.
Geetest
- domain
- gt
- challenge
- api_server (optional)
- user_agent (optional)
- proxy (optional)
d = {'domain': 'DOMAIN_HERE', 'gt': 'GT_HERE', 'challenge': 'CHALLENGE_HERE'}
# d['api_server'] = 'GT_DOMAIN_HERE' # optional
# d['user_agent'] = 'your user agent'
# d['proxy'] = 'user:pass@123.45.67.89:3031' # optional
captcha_id = bcs.submit_geetest(d)
Use captcha_id to retrieve solution
for geetest
GeetestV4
- domain
- captchaid
- user_agent (optional)
- proxy (optional)
Important: This is not the captchaid that's in our system that you receive while submitting a captcha. Gather this from HTML source of page with geetestv4 captcha, inside the <script>
tag you'll find a link that looks like this: https://i.imgur.com/XcZd47y.png
d = {'domain': 'https://example.com', 'captchaid': '647f5ed2ed8acb4be36784e01556bb71'}
# d['user_agent'] = 'your user agent'
# d['proxy'] = 'user:pass@123.45.67.89:3031' # optional
captcha_id = bcs.submit_geetest_v4(d)
Use captcha_id received from service to retrieve solution
for geetestv4
Capy
- page_url
- site_key
- user_agent (optional)
- proxy (optional)
captcha_id = bcs.submit_capy({'page_url': 'PAGE_URL_HERE', 'site_key': 'SITEKEY_HERE'}) # , 'proxy': 'user:pass@123.45.67.89:3031', 'user_agent': 'your user agent'})
Use captcha_id to retrieve solution
for capy
hCaptcha
- page_url
- site_key
- invisible (optional)
- payload (optional)
- domain (optional)
- user_agent (optional)
- proxy (optional)
d = {'page_url': 'PAGE_URL_HERE', 'site_key': 'SITEKEY_HERE'}
# d['invisible'] = True
# d['payload'] = {'rqdata': 'gather from page source, unique with each submission'}
# d['domain'] = 'hcaptcha.com' # optional
# d['user_agent'] = 'your user agent' # optional
# d['proxy'] = 'user:pass@123.45.67.89:3031' # optional
captcha_id = bcs.submit_hcaptcha(d)
Use captcha_id to retrieve solution
for hCaptcha
FunCaptcha (Arkose Labs)
- page_url
- s_url
- site_key
- data (optional)
- user_agent (optional)
- proxy (optional)
captcha_id = bcs.submit_funcaptcha(
{'page_url': 'https://abc.com', 'site_key': '11111111-1111-1111-1111-111111111111',
's_url': 'https://api.arkoselabs.com'}
# 'data': '{"x":"y"}', # optional
# 'user_agent': 'your user agent',
# 'proxy': 'user:pass@123.45.67.89:3031'
)
Turnstile (Cloudflare)
- page_url
- site_key
- action (optional)
- cdata (optional)
- domain (optional)
- user_agent (optional)
- proxy (optional)
d = {'page_url': 'PAGE_URL_HERE', 'site_key': 'SITEKEY_HERE'}
# d['action'] = 'taken from page source, optional'
# d['cdata'] = 'taken from page source, optional'
# d['domain'] = 'challenges.cloudflare.com' # optional
# d['user_agent'] = 'your user agent' # optional
# d['proxy'] = 'user:pass@123.45.67.89:3031'
captcha_id = bcs.submit_turnstile(d)
Task
- template_name
- page_url
- variables
data = {
'template_name': 'Login test page',
'page_url': 'https://bestcaptchasolver.com/automation/login',
'variables': {"username": "xyz", "password": "0000"},
# 'proxy': '126.45.34.53:345', # or user:pass@123.45.67.89:3031
# 'user_agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Firefox/60.0', # optional
}
captcha_id = bcs.submit_task(data)
Update task variables while it is being solved by the worker. Useful when dealing with data / variables, of which value you don't know, only after a certain step or action of the task. For example, in websites that require 2 factor authentication code.
When the task (while running on workers machine) is getting to an action defined in the template, that requires a variable, but variable was not set with the task submission, it will wait until the variable is updated through push.
The bcs.task_push_variables(captcha_id, push_variables)
method can be used as many times as it is needed.
bcs.task_push_variables(captcha_id, dict(tfa_code=57))
Retrieve captcha response (all captchas)
image_text = bcs.retrieve(captcha_id)['text']
gresponse = bcs.retrieve(recaptcha_id)['gresponse']
solution = bcs.retrieve(captcha_id)['solution']
If submitted with proxy, get proxy status
proxy_status = bcs.retrieve(recaptcha_id)['proxy_status']
Set captcha bad
When a captcha was solved wrong by our workers, you can notify the server with it's ID, so we know something went wrong.
bcs.set_captcha_bad(captcha_id)
Check main.py
API library is licensed under the MIT License
More details about the server-side API can be found here
captcha, bypasscaptcha, decaptcher, decaptcha, 2captcha, deathbycaptcha, anticaptcha, bypassrecaptchav2, bypassnocaptcharecaptcha, bypassinvisiblerecaptcha, captchaservicesforrecaptchav2, recaptchav2captchasolver, googlerecaptchasolver, recaptchasolverpython, recaptchabypassscript, bestcaptchasolver