CAPTCHA solving API

Alibaba Cloud CAPTCHA can appear as a click-through, slider, puzzle, or other risk-adaptive widget; it is not Alibaba's separate page-level shield. The returned token belongs to the widget verification flow; do not treat it as a page-access cookie.

CaptchaFox is a site-key widget; identify its script or widget configuration before copying the key. It returns a widget token, not a page-access cookie. Do not reuse a reCAPTCHA or Turnstile key simply because the DOM also says sitekey.

Capy Puzzle CAPTCHA uses a site key and may point to a deployment-specific API server. The response has captchakey, challengekey, and answer fields that must be read together; it is not a single token.

Friendly Captcha is a proof-of-work widget with its own site key and optional client library endpoint. It returns a widget token. It is not ALTCHA, MTCaptcha, or an invisible reCAPTCHA merely because the interaction is low-friction.

Arkose Labs FunCaptcha is commonly exposed through an Arkose script or a data-pkey/public-key value and may present interactive 3D or visual tasks. It returns the Arkose session token, commonly consumed through an fc-token field. It is not a reCAPTCHA Enterprise payload.

GeeTest uses a CAPTCHA ID/site key and can render adaptive, slide, click, or other challenge forms. The result is a five-field verification set: captcha_id, lot_number, pass_token, gen_time, and captcha_output, not one generic token.

hCaptcha is an h-captcha widget or explicit hcaptcha.render call with its own data-sitekey. The API returns the hCaptcha response in `gRecaptchaResponse`. The field name does not make the widget Google reCAPTCHA.

KeyCAPTCHA is an older widget with several signed page values, not a single site-key integration. It returns the KeyCAPTCHA token; none of the four identifiers can be replaced with a reCAPTCHA-style data-sitekey.

Lemin CAPTCHA can render boxed or invisible modes and uses a site key plus an optional container and API server. It returns answer and challenge_id together. Do not treat it as the single token returned by MTCaptcha or Turnstile.

MTCaptcha is its own widget, commonly configured with an MTCaptcha site key; it is not Yandex SmartCaptcha despite the similar name. It returns an MTCaptcha token. Yandex SmartCaptcha loads Yandex Cloud scripts and uses smart-token; the site keys are not interchangeable.

Prosopo Procaptcha is a site-key widget that can present proof-of-work or visual challenges according to its deployment. It returns a Procaptcha token. A data-sitekey attribute alone does not make it reCAPTCHA, hCaptcha, or Turnstile.

The reCAPTCHA grid task solves an already extracted image question; it does not navigate a page or generate a reCAPTCHA site token. It returns selected object positions. Use `recaptcha_v2` for the complete widget token workflow.

reCAPTCHA v2 is the classic checkbox or invisible v2 widget loaded through api.js and configured with a site key. It returns gRecaptchaResponse for the target form. v3 is score-based and Enterprise loads enterprise.js with Enterprise-specific context.

reCAPTCHA v2 Enterprise is a checkbox or policy-driven Enterprise key loaded through enterprise.js and grecaptcha.enterprise. It still returns gRecaptchaResponse, but Enterprise identifiers and payloads must not be mixed with classic v2 api.js values or score-based v3 keys.

reCAPTCHA v3 is score-based and normally runs through grecaptcha.execute without showing the v2 checkbox. EverySolve returns gRecaptchaResponse, not Google's assessment score. The protected site's backend evaluates the token and action.

reCAPTCHA v3 Enterprise uses an Enterprise score-based key with enterprise.js and grecaptcha.enterprise.execute. It returns gRecaptchaResponse for Enterprise assessment. It has no v2 checkbox contract and is not classic v3 api.js merely because both are invisible.

Tencent Tendi CAPTCHA uses a captchaAppId/site key and can return a ticket plus randstr after its interactive challenge. The `ticket` and `randstr` parameters must be consumed together. They are not a reCAPTCHA response token or a page-access cookie.

Cloudflare Turnstile is an embedded widget configured with a Turnstile site key. It may be managed, non-interactive, or invisible. It returns a token for `cf-turnstile-response`. A Cloudflare challenge page instead blocks navigation and returns clearance cookie data.

NetEase Yidun, also called NECaptcha, initializes with its own captchaId and can present slider, click, or risk-adaptive challenges. It returns a Yidun token. Its captchaId and callback values are not Tencent ticket/randstr or GeeTest's five-field result.

This task transcribes an audio CAPTCHA payload; it does not fetch audio from a page or solve an image challenge. It returns text. Use image for characters in a still image and text_captcha for a written question.

This is the image-classification half of an AWS WAF CAPTCHA, using challenge images plus the displayed instruction. It returns click positions in box, not an AWS WAF passage cookie. The caller still controls the browser interaction.

BotDeflector here is an image-selection task: the submitted body is analyzed for point locations. It returns points, not a form token or browser cookie, so the caller remains responsible for applying the positions.

A bounding-box task asks for rectangles around matching objects in one source image. It returns bounding_boxes with rectangle geometry. Use coordinates for click points and recaptcha_grid for tile or object selections.

A coordinate task represents one or more click locations in an image, independent of any single CAPTCHA vendor. It returns coordinate points rather than rectangles, text, or a drag distance. The caller applies those points to the original image dimensions.

The image task performs character recognition on an image payload; it is not a branded widget integration. It returns recognized text. Use audio for sound, text_captcha for a written question, and coordinate or box tasks for spatial answers.

A rotation task receives one image that must be turned upright; it is a general visual task. It returns angle as a number. Use slider for a drag distance, coordinates for clicks, and bounding_box for rectangles.

The SHEIN task is a site-specific icon-location model rather than the general bounding-box task. It returns rects for the matching regions. Images from other sites should use a suitable generic point or box task instead.

A slider task uses a background image and a separate puzzle piece to calculate how far the piece should move. It returns distance in pixels, not a coordinate, percentage, token, or drag trajectory. The caller creates the interaction path.

The Temu task matches an ordered parts array against a main challenge image. It returns coordinates in the same order as the candidate images in parts. It is not the single-image coordinate task and does not return a site token.

A text CAPTCHA is a written question-and-answer prompt, often supplied by a forum or CMS rather than a hosted widget vendor. It returns answer text. Use image for pictured characters and audio for a spoken challenge.