Claude Vision (in-house CAPTCHA) integration

In-house AI CAPTCHA solving via Claude Vision – image grids, OCR, object ID. No third-party CAPTCHA service.

Help CentreConnectors

Overview

Captcha via Claude Vision is Automize's in-house CAPTCHA solver – uses Anthropic's Claude Vision model to solve image-based CAPTCHAs (recaptcha v2 grids, image OCR, object identification). No external third-party fees; uses your Anthropic API key.

  • Regions: Wherever Anthropic API is available (global).
  • Plans: Pay-per-Claude-API-call. Cheaper than dedicated CAPTCHA services for image-based challenges; not applicable for token-based challenges (reCAPTCHA v3, hCaptcha, Turnstile – use 2Captcha/Capsolver for those).

Setting up the connection

  1. Uses your existing Anthropic API key – same as the anthropic connector.
  2. In Automize → Settings → Connectors → Captcha (Claude Vision). Paste Anthropic API key.

Find your credentials at https://console.anthropic.com/settings/keys.


Rate limits

Bound by your Anthropic tier (~5-15 RPS on tier 1, scales with usage).


Data model

Image-based reasoning:

    base64 image + question → Claude vision endpoint → coordinates / text / labels

Available operations

3 operations available. Click any row to jump to its detail.

OperationCategoryWhat it does
Identify Objects in ImageGeneralGeneric vision query – list objects present, useful for atypical CAPTCHA flows
Solve Image Grid CAPTCHAGeneralPick which tiles in a 3x3 or 4x4 image grid match the prompt (e.g. 'traffic lights')
Solve OCR / Distorted-Text CAPTCHAGeneralRead the text shown in a distorted-text CAPTCHA image

Operations

Identify Objects in Image

Generic vision query – list objects present, useful for atypical CAPTCHA flows

ParameterRequiredTypeDescription
Image Path image_pathYesstringPath to the image
Question questionYesstringNatural-language question to ask about the image

Solve Image Grid CAPTCHA

Pick which tiles in a 3x3 or 4x4 image grid match the prompt (e.g. 'traffic lights')

ParameterRequiredTypeDescription
Image Path image_pathYesstringPath to a screenshot of the full grid
Prompt promptYesstringChallenge text, e.g. 'Select all images with traffic lights'
Grid Size grid_sizeNointegerGrid dimension (3 for 3x3, 4 for 4x4)

Solve OCR / Distorted-Text CAPTCHA

Read the text shown in a distorted-text CAPTCHA image

ParameterRequiredTypeDescription
Image Path image_pathYesstringPath to the CAPTCHA image
Hint hintNostringOptional context, e.g. 'lowercase letters and digits, 6 chars'

FAQ

When to use Claude Vision vs 2Captcha?
Claude Vision: image-based + custom CAPTCHA logic, lower cost. 2Captcha: token-based (reCAPTCHA v2/v3, hCaptcha, Turnstile) where you need a server-issued token.
Accuracy?
High on image grids + OCR. Lower on highly-distorted modern CAPTCHAs designed to defeat AI. Test on your specific target site.

Related connectors

See it working on your own data

Everything documented here ships with the platform – try the document tools free, or go live in 7 days.