Captcha types

reCAPTCHA v2#

Types RecaptchaV2TaskProxyless and RecaptchaV2Task.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the captcha is located
websiteKey	yes	string	Value of the data-sitekey attribute of the reCAPTCHA widget
isInvisible	no	bool	true for invisible reCAPTCHA. Default is false
recaptchaDataSValue	no	string	Value of the data-s parameter, which may be required on Google services
userAgent	no	string	User-Agent of the browser used to load the captcha
cookies	no	string	Cookies in the format name1=value1; name2=value2, if the session is important for solving
apiDomain	no	string	Domain used to load reCAPTCHA: google.com or recaptcha.net. Default is google.com

Only for RecaptchaV2Task (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Task object examples

// RecaptchaV2TaskProxyless
{
  "type": "RecaptchaV2TaskProxyless",
  "websiteURL": "https://example.com/login",
  "websiteKey": "6Le-xxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "isInvisible": false
}

// RecaptchaV2Task (your proxy)
{
  "type": "RecaptchaV2Task",
  "websiteURL": "https://example.com/login",
  "websiteKey": "6Le-xxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080,
  "proxyLogin": "user",
  "proxyPassword": "password",
  "userAgent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) ..."
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "gRecaptchaResponse": "03AGdBq..."
  }
}

You need to pass the solution.gRecaptchaResponse value to the reCAPTCHA widget or the g-recaptcha-response form parameter.

reCAPTCHA v2 Enterprise#

Types RecaptchaV2EnterpriseTaskProxyless and RecaptchaV2EnterpriseTask.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the captcha is placed
websiteKey	yes	string	Value of the data-sitekey attribute of the reCAPTCHA Enterprise widget
enterprisePayload	no	object	Additional parameters of the grecaptcha.enterprise.render call, if the site passes them, for example the s value
isInvisible	no	bool	true for invisible reCAPTCHA. Default is false
apiDomain	no	string	Domain from which reCAPTCHA is loaded: google.com or recaptcha.net. Default is google.com
userAgent	no	string	User-Agent of the browser used to load the captcha
cookies	no	string	Cookies in the format name1=value1; name2=value2, if the session is important for solving

Only for RecaptchaV2EnterpriseTask (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Examples of the task object

// RecaptchaV2EnterpriseTaskProxyless
{
  "type": "RecaptchaV2EnterpriseTaskProxyless",
  "websiteURL": "https://example.com/login",
  "websiteKey": "6Le-xxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "isInvisible": false
}

// RecaptchaV2EnterpriseTask (your proxy)
{
  "type": "RecaptchaV2EnterpriseTask",
  "websiteURL": "https://example.com/login",
  "websiteKey": "6Le-xxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "isInvisible": false,
  "userAgent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) ...",
  "cookies": "foo=bar; baz=1",
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080,
  "proxyLogin": "user",
  "proxyPassword": "password"
}

Example of getTaskResult response

While the task is being processed:

{
  "errorId": 0,
  "status": "processing"
}

After successful solving:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "gRecaptchaResponse": "03AGdBq2kX9Lm...QpZr8vN3sT"
  }
}

The solution.gRecaptchaResponse value is passed in the same way as a regular reCAPTCHA v2 token, to the widget or to the g-recaptcha-response parameter of the form.

The Enterprise method uses the reCAPTCHA Enterprise API. If the site passes additional parameters through enterprisePayload, include them in the request.

reCAPTCHA v3#

Type RecaptchaV3TaskProxyless. A proxy is not required for v3. Tasks are solved from the service IP addresses.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page with the captcha
websiteKey	yes	string	Site key of the reCAPTCHA v3 widget
minScore	yes	float	Required score. Supported values are 0.3, 0.7, and 0.9
pageAction	no	string	Value of the action parameter that the site sets when calling grecaptcha.execute()
isEnterprise	no	bool	true if the site uses reCAPTCHA v3 Enterprise
apiDomain	no	string	Domain from which the reCAPTCHA script is loaded: google.com or recaptcha.net. Default is google.com

Task object example

// RecaptchaV3TaskProxyless
{
  "type": "RecaptchaV3TaskProxyless",
  "websiteURL": "https://example.com/login",
  "websiteKey": "6Le-xxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "minScore": 0.3,
  "pageAction": "login",
  "isEnterprise": false
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "gRecaptchaResponse": "03AGdBq..."
  }
}

You use the received token (solution.gRecaptchaResponse) just like a regular reCAPTCHA v3 token.

Finding the reCAPTCHA sitekey and callback function on the page#

If there are multiple reCAPTCHA widgets on the page or the callback is not directly visible in the markup, you can find the sitekey and callback function via the ___grecaptcha_cfg object, where Google stores the configuration of rendered widgets. Run the following code in the developer console or embed it in your automation script. It will return an array with the found reCAPTCHA clients along with their sitekey, pageurl, and callback reference.

function findRecaptchaClients() {
  // eslint-disable-next-line camelcase
  if (typeof (___grecaptcha_cfg) !== 'undefined') {
    // eslint-disable-next-line camelcase, no-undef
    return Object.entries(___grecaptcha_cfg.clients).map(([cid, client]) => {
      const data = { id: cid, version: cid >= 10000 ? 'V3' : 'V2' };
      const objects = Object.entries(client).filter(([_, value]) => value && typeof value === 'object');

      objects.forEach(([toplevelKey, toplevel]) => {
        const found = Object.entries(toplevel).find(([_, value]) => (
          value && typeof value === 'object' && 'sitekey' in value && 'size' in value
        ));
     
        if (typeof toplevel === 'object' && toplevel instanceof HTMLElement && toplevel['tagName'] === 'DIV') {
          data.pageurl = toplevel.baseURI;
        }
        
        if (found) {
          const [sublevelKey, sublevel] = found;

          data.sitekey = sublevel.sitekey;
          const callbackKey = data.version === 'V2' ? 'callback' : 'promise-callback';
          const callback = sublevel[callbackKey];

          if (!callback) {
            data.callback = null;
            data.function = null;
          } else {
            data.function = callback;
            const keys = [cid, toplevelKey, sublevelKey, callbackKey]
              .map((key) => `['${key}']`)
              .join('');

            data.callback = `___grecaptcha_cfg.clients${keys}`;
          }
        }
      });

      return data;
    });
  }

  return [];
}

Usage:

let res = findRecaptchaClients()
console.log(res)

For each found client, data.version indicates whether it is v2 or v3. data.sitekey is the value for the websiteKey field in the createTask request. data.callback is the path to the callback function that needs to be executed with the token after receiving the solution.

Cloudflare Turnstile#

Types TurnstileTaskProxyless and TurnstileTask.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the Turnstile widget is located
websiteKey	yes	string	Value of the data-sitekey attribute of the Turnstile widget
action	no*	string	Value of the action parameter from the turnstile.render call. Required for Cloudflare Challenge pages
data	no*	string	Value of the cData parameter from the turnstile.render call. Required for Cloudflare Challenge pages
pagedata	no*	string	Value of the chlPageData parameter from the turnstile.render call. Required for Cloudflare Challenge pages
userAgent	no*	string	Current browser User-Agent. Required for Cloudflare Challenge pages

Only for TurnstileTask (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Task object examples

// TurnstileTaskProxyless
{
  "type": "TurnstileTaskProxyless",
  "websiteURL": "https://example.com/login",
  "websiteKey": "0x4AAAAAAAxxxxxxxxxxxxxxxx"
}

// TurnstileTask (your proxy)
{
  "type": "TurnstileTask",
  "websiteURL": "https://example.com/login",
  "websiteKey": "0x4AAAAAAAxxxxxxxxxxxxxxxx",
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "token": "0.zxcv..."
  }
}

You need to pass the received token (solution.token) to the widget callback function or the cf-turnstile-response field, depending on how you integrate Turnstile on your site.

Complex case: Cloudflare Challenge page#

A separate and more complex situation is the Cloudflare Challenge page. Besides websiteKey, additional parameters cData, chlPageData, and action are passed to the widget. In this case, you need to intercept the turnstile.render call to extract these values. You also need to intercept the callback function definition.

To intercept the parameters, embed the following JavaScript on the page before the Cloudflare Turnstile widget itself loads:

const i = setInterval(() => {
  if (window.turnstile) {
    clearInterval(i)

    window.turnstile.render = (a, b) => {
      let p = {
        type: "TurnstileTaskProxyless",
        websiteKey: b.sitekey,
        websiteURL: window.location.href,
        data: b.cData,
        pagedata: b.chlPageData,
        action: b.action,
        userAgent: navigator.userAgent
      }

      console.log(JSON.stringify(p))
      window.tsCallback = b.callback

      return 'foo'
    }
  }
}, 10)

The script replaces turnstile.render with a custom function. It intercepts the configuration object b, in which Cloudflare passes sitekey, cData, chlPageData, action, and callback. It logs the task object for createTask and saves a reference to the original callback function in window.tsCallback to call it later with the token.

An alternative approach is to intercept the network request to the api.js script and replace it with your own script. This script returns the required parameters and makes the callback function globally accessible.

When the solution is received from getTaskResult, execute the saved callback function by passing the token as an argument:

window.tsCallback('TOKEN_FROM_SOLUTION');

Keep the browser User-Agent unchanged when invoking the callback. The response may contain solution.userAgent; do not switch the browser to this value.

Yandex SmartCaptcha#

There are two methods for solving Yandex SmartCaptcha: token-based and image-based.

• YandexSmartCaptchaTaskProxyless — token-based solving using our proxies
• YandexSmartCaptchaTask — token-based solving using your proxies
• CoordinatesTask — image-based solving, returns click coordinates

Token-based solving#

Types YandexSmartCaptchaTaskProxyless and YandexSmartCaptchaTask.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the captcha is loaded
websiteKey	yes	string	The sitekey value from the page code or captcha iframe
userAgent	no	string	User-Agent under which the captcha is solved
cookies	no	string	Cookies in the format name1=value1;name2=value2

Only for YandexSmartCaptchaTask (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, https, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Task object examples

// YandexSmartCaptchaTaskProxyless
{
  "type": "YandexSmartCaptchaTaskProxyless",
  "websiteURL": "https://example.com/login",
  "websiteKey": "Y5Lh0ti..."
}

// YandexSmartCaptchaTask (your proxy)
{
  "type": "YandexSmartCaptchaTask",
  "websiteURL": "https://example.com/login",
  "websiteKey": "Y5Lh0ti...",
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "token": "dV9xNjYyNTU3NjkxO4k9OTQuNVMuMjkuMjM9..."
  }
}

solution.token should be used in the smart-token field or passed to your site's backend, the same way as a manually solved captcha.

Image to Text#

Type ImageToTextTask. No proxy is required. The captcha image itself is submitted directly.

Parameter	Required	Type	Description
body	yes	string	The captcha image, Base64-encoded. Data-URI format is also supported
phrase	no	bool	true if the answer should contain at least two words separated by a space. Default is false
case	no	bool	true if the result is case-sensitive. Default is false
numeric	no	int	Character set hint: 0 — no preference, 1 — numbers only, 2 — letters only, 3 — numbers or letters only, 4 — must contain both numbers and letters
math	no	bool	true if the captcha requires a calculation. Default is false
minLength	no	int	Minimum answer length. 0 means no minimum
maxLength	no	int	Maximum answer length. 0 means no maximum
comment	no	string	A hint for the worker solving the image
imgInstructions	no	string	An optional instruction image, Base64-encoded, shown to the worker alongside the captcha image

Task object example

{
  "type": "ImageToTextTask",
  "body": "iVBORw0KGgoAAAANSUhEUgAA...",
  "numeric": 1,
  "minLength": 4,
  "maxLength": 6
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "text": "aB3fX9"
  }
}

solution.text is the recognized text. Submit it to the target form field the same way you would a manually entered answer.

Coordinates#

Use CoordinatesTask for generic click-on-image captchas.

Parameter	Required	Description
body	yes	Captcha image encoded as Base64
comment	no	Text instruction for the worker
imgInstructions	no	Optional instruction image encoded as Base64
minClicks	no	Minimum number of clicks
maxClicks	no	Maximum number of clicks

The solution contains a coordinates array with click positions.

GeeTest v3#

Types GeeTestTaskProxyless and GeeTestTask. The same task types are used for GeeTest v4. The version is selected with the version field. The default version is 3.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the GeeTest widget is located
version	no	int	GeeTest version: 3 or 4. Default is 3
gt	yes	string	Public gt value of the GeeTest widget
challenge	yes	string	Session-specific challenge value found on the page. Get a new value for each task
geetestApiServerSubdomain	no	string	Custom GeeTest API server subdomain, if the page uses one
initParameters	no	object	Additional parameters passed to the GeeTest initialization call
userAgent	no	string	User-Agent of the browser used to load the captcha
risk_type	no	string	Dynamic risk_type value included in the captcha loading request

Only for GeeTestTask (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Task object examples

// GeeTestTaskProxyless
{
  "type": "GeeTestTaskProxyless",
  "websiteURL": "https://example.com/login",
  "gt": "f2ae6cadcf7886856696c46d84d109d1",
  "challenge": "12345678abc90123d45678e90123f45g6"
}

// GeeTestTask (your proxy)
{
  "type": "GeeTestTask",
  "websiteURL": "https://example.com/login",
  "gt": "f2ae6cadcf7886856696c46d84d109d1",
  "challenge": "12345678abc90123d45678e90123f45g6",
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "challenge": "12345678abc90123d45678e90123f45g6",
    "validate": "0a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p",
    "seccode": "0a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p|jordan"
  }
}

Pass solution.validate and solution.seccode into the page's GeeTest callback according to the integration used by the target site.

GeeTest v4#

Same types as GeeTest v3, GeeTestTaskProxyless and GeeTestTask. Set version to 4 to use GeeTest v4.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the GeeTest v4 widget is located
version	yes	int	Must be 4
initParameters	yes	object	Must contain captcha_id, the widget's static captcha identifier
userAgent	no	string	User-Agent of the browser used to load the captcha
risk_type	no	string	Dynamic risk_type value included in the captcha loading request

Only for GeeTestTask (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Task object examples

// GeeTestTaskProxyless (v4)
{
  "type": "GeeTestTaskProxyless",
  "websiteURL": "https://example.com/login",
  "version": 4,
  "initParameters": {
    "captcha_id": "e392e65f912c780f2c3ebac7702651de"
  }
}

// GeeTestTask (v4, your proxy)
{
  "type": "GeeTestTask",
  "websiteURL": "https://example.com/login",
  "version": 4,
  "initParameters": {
    "captcha_id": "e392e65f912c780f2c3ebac7702651de"
  },
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "captcha_id": "e392e65f912c780f2c3ebac7702651de",
    "lot_number": "...",
    "pass_token": "...",
    "gen_time": "...",
    "captcha_output": "..."
  }
}

solution.lot_number, pass_token, gen_time, and captcha_output together form the GeeTest v4 result object. Pass them into the page's GeeTest callback as-is.

Tencent#

Types TencentTaskProxyless and TencentTask.

Parameter	Required	Type	Description
websiteURL	yes	string	Full URL of the page where the captcha is loaded
appId	yes	string	The value of the appId parameter found in the page source code
captchaScript	no	string	Captcha script URL from the page source code. Default: https://turing.captcha.qcloud.com/TCaptcha.js

Only for TencentTask (solving via your proxy), additionally:

Parameter	Required	Type	Description
proxyType	yes	string	http, socks4, or socks5
proxyAddress	yes	string	Proxy IP address or hostname
proxyPort	yes	int	Proxy port
proxyLogin	no	string	Login for proxy authorization
proxyPassword	no	string	Password for proxy authorization

Task object examples

// TencentTaskProxyless
{
  "type": "TencentTaskProxyless",
  "websiteURL": "https://example.com/login",
  "appId": "190014885"
}

// TencentTask (your proxy)
{
  "type": "TencentTask",
  "websiteURL": "https://example.com/login",
  "appId": "190014885",
  "proxyType": "http",
  "proxyAddress": "1.2.3.4",
  "proxyPort": 8080
}

getTaskResult response example

While the task is processing:

{
  "errorId": 0,
  "status": "processing"
}

Once solved:

{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "appid": "190014885",
    "ret": 0,
    "ticket": "tr0344YjJASGmJGtohyWS_y6tJKiqVPIdFgl87vWlVaQoueR8D6DH28go-i-VjeassM31SXO7D0*",
    "randstr": "@KVN"
  }
}

solution.appid, ret, ticket, and randstr together form the Tencent result object. Pass them into the page's captcha callback as-is.

Python SDK#

All listed task types are available in captcha_solver_api.tasks. Parameters preserve API casing: websiteURL, websiteKey, enterprisePayload, and imgInstructions; use pagedata and risk_type exactly as written. None fields are omitted, while False and 0 are preserved. Pass language_pool to the client or task-creation method, outside the task object. Examples for each type.