Skip to main content

CAPTCHA guide

ATB Captcha (atbCAPTCHA)

Risk-adaptive slider, click or invisible check, validated in several stages.

Difficulty: Medium
Open source
Exploded concept drawing of how the ATB CAPTCHA challenge is built

atbCAPTCHA is a behavioural risk-control product. Depending on its risk score it shows a click, slider or image-alignment task, or nothing at all. It is found in finance, telecoms, online retail and IoT, and it checks traffic in several stages instead of relying on one puzzle.

Challenge typesLink to Challenge types

  • Slider: line up a piece or a cut-out
  • Image click: click or align the requested images
  • Invisible: a token is issued silently to low-risk sessions
  • Multi-stage validation: more checks run after the visible task

How to identifyLink to How to identify

  • Global appId and apiServer values in the page source
  • apiServer frequently set to https://cap.aisecurius.com
  • SDK setup with a success callback: as.Captcha(element, {appId, apiServer, success})
  • A token handed to that callback, often a short prefix followed by a colon
  • atbCAPTCHA branding, or a widget served from the apiServer domain
  • A hidden form field or JS callback that receives the final token

How it worksLink to How it works

  1. The page starts the widget with appId and apiServer.
  2. The client asks for a challenge, and atbCAPTCHA picks its type from the risk score.
  3. The visible task, or the invisible check, produces a token.
  4. The token reaches the site's success callback and is sent to the backend.
  5. atbCAPTCHA validates it along with the IP, User-Agent, timing and behaviour.

Free bypass methodsLink to Free bypass methods

Method What it does Success Rate Link
DrissionPage Runs the widget in real Chromium and catches the callback token 70-85% g1879/DrissionPage
OpenCV Finds the slider gap via template matching or edges 70-90% (visible slider) opencv/opencv-python
ddddocr OCR and object detection for image tasks 65-85% sml2h3/ddddocr
YOLOv8 Detects target images in click and align tasks 70-90% (with a trained model) ultralytics/ultralytics
Patchright / Camoufox Stealth browsers that keep the risk score down 60-85% Kaliiiiiiiiii-Vinyzu/patchright-python
undetected-chromedriver Selenium with detection patches 60-80% ultrafunkamsterdam/undetected-chromedriver

Practical approachLink to Practical approach

  1. Note the appId, the apiServer and the page URL.
  2. Load the page in a real browser on a clean residential IP.
  3. When the check is invisible, simply wait for the success callback.
  4. When a slider shows up, solve it with OpenCV and drag the way a person would.
  5. When an image task appears, find the requested pictures and click them in order.
  6. Take the token from the callback or the hidden field and submit right away.

Practical tipsLink to Practical tips

  • The token is linked to the IP and User-Agent that earned it. Solve and submit through the same proxy.
  • Validation has several stages: a correct puzzle answer still fails if the fingerprint looks wrong.
  • Datacenter IPs are usually rated high-risk. Use residential addresses.
  • Reloading the widget after solving can void the token.
  • Risk-adaptive systems watch how you interact, not just the final answer, so keep mouse paths natural.

Skip the solver: let BrowserSolver handle ATB CAPTCHA

Website AntiBot runs a real browser session with matching TLS, headers, cookies and IP, clears the challenge, and returns the page. You send a URL, you get HTML back.

For education and authorized testing only. Respect the law, privacy, and the terms of the sites you work with. Difficulty labels are editorial estimates. Adapted from the CC0 community field guide on GitHub.