CaptchaSonic icon

CaptchaSonic

AI-powered CAPTCHA solving — supports reCAPTCHA, popularcaptcha, Turnstile, GeeTest, DataDome, AWS WAF, and more

Actions21

Recognition → Vision Engine

AI-generated

Summary

Sends a CAPTCHA image and/or question to CaptchaSonic's AI Vision Engine task and returns the solver's inline response for complex image challenges.

Inputs

  • Task Type (required) — Fixed to VisionEngine; sent as the task's type.
  • Image (Base64) — Base64-encoded CAPTCHA image. Sent as the first entry of the task's queries array.
  • Question — Instruction text for the challenge (e.g. "Select all images with cars"). Sent as question.
  • Website URL — Page where the CAPTCHA appears. Sent as websiteURL if non-empty.
  • Website Key — Sitekey of the CAPTCHA widget, if applicable. Sent as both websiteKey and websiteKEY.
  • Body — Raw body data. If it parses as a JSON object, its fields are merged into the task (useful for extras such as questionType, screenshot, choices, examples) and can override earlier fields. Otherwise it is sent as a raw string in body.

Output shape

One output item per input item: { success: true, data: }, paired to the input item.

Recognition tasks return the result directly from the createTask call, with no polling. The exact fields in data depend on the API. If the API returns a non-zero errorId, the node throws, or with Continue On Fail it outputs { error: message } for that item. Requires the captchaSonicApi credential (API key).

Examples

Example 1: Solve an image-selection challenge from a captured screenshot

Resource: Recognition, Operation: Vision Engine; Image (Base64) = {{$json.screenshot}}; Question = "Select all images with traffic lights".

Example 2: Pass extra API-specific fields for a custom challenge

Image (Base64) set; Body = {"questionType":"grid","choices":["A","B"]} to be merged into the task.

Links

Discussion