On this page

Docs · reCAPTCHA

reCAPTCHA v3 & v2

One call returns a reCAPTCHA v3 or v2 token: send your API key, the page URL and its sitekey. For v3, also send the page's action (pageAction) and its reCAPTCHA script URL (apiJs).

  • Endpoint POST /recaptcha/getToken
  • Token 120 s, works once
  • API as JSON

Quick start: reCAPTCHA v3

Three steps, no polling loop: getToken waits and returns the token. Copy, paste, replace the example values.

1

Get a key

API keys are issued in our Telegram bot @mintprocheckerbot. Solver access is opening soon — follow @MINTBotUpdates for the launch.

Send your key as clientKey in every request, next to task. Keep it on your server.

2

Send this request

POST it to /recaptcha/getToken. The reply comes when the token is ready.

cURL
curl -X POST "https://solver.mintbin.xyz/recaptcha/getToken" \
  -H "Content-Type: application/json" \
  -d '{
    "clientKey": "YOUR_API_KEY",
    "task": {
      "type": "ReCaptchaV3HSTaskProxyLess",
      "websiteURL": "https://example.com/my-account",
      "websiteKey": "6Lxxxx...",
      "pageAction": "login",
      "apiJs": "https://www.google.com/recaptcha/api.js?render=6Lxxxx..."
    }
  }'
  • websiteURL: the address of the page that shows the captcha.
  • websiteKey: the 6L… key, usually the render= value in the apiJs URL. If apiJs has render=explicit (or no render=), search the page source for the 6L… key and send that.
  • pageAction: the action name the page sends with reCAPTCHA, e.g. login. To find it, search the page's scripts (DevTools → Sources, Ctrl+Shift+F) for action:.
  • How to find apiJs: open DevTools → Network, type recaptcha in the filter, reload the page, and copy the full api.js (or enterprise.js) URL, query string included.
3

Use the token

The token is in solution.gRecaptchaResponse. Send it the way the page does (usually the g-recaptcha-response form field) and submit within expires_in_s seconds (at most 120). Each token works once, and only on the site in websiteURL. Each call waits up to 60 s by default.

Expected reply

JSON
{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "gRecaptchaResponse": "03AGdBq26...long-opaque-token..."
  },
  "taskId": "3f2a...",
  "elapsed_s": 8.3,
  "expires_in_s": 119
}

Submit it with the form

# TOKEN = solution.gRecaptchaResponse from the getToken reply
curl -X POST "https://example.com/my-account" \
  --data-urlencode "email=you@example.com" \
  --data-urlencode "g-recaptcha-response=$TOKEN"

Other task types →

Quick start: reCAPTCHA v2

The “I'm not a robot” checkbox, including image challenges. Same three steps; pageAction and apiJs are not used.

1

Get a key

API keys are issued in our Telegram bot @mintprocheckerbot. Solver access is opening soon — follow @MINTBotUpdates for the launch.

Send your key as clientKey in every request, next to task. Keep it on your server.

2

Send this request

cURL
curl -X POST "https://solver.mintbin.xyz/recaptcha/getToken" \
  -H "Content-Type: application/json" \
  -d '{
    "clientKey": "YOUR_API_KEY",
    "task": {
      "type": "ReCaptchaV2TaskProxyLess",
      "websiteURL": "https://example.com/login",
      "websiteKey": "6Lxxxx..."
    }
  }'
  • websiteURL: the address of the page that shows the checkbox.
  • websiteKey: the checkbox's data-sitekey (also the k= value in its iframe URL).
3

Use the token

Put solution.gRecaptchaResponse in the form's g-recaptcha-response field and submit within expires_in_s seconds (at most 120). Each token works once.

JSON
{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "gRecaptchaResponse": "03AFcWeA5...long-opaque-token..."
  },
  "taskId": "9c41...",
  "elapsed_s": 18.4,
  "expires_in_s": 119
}

Code to submit the token ↑

An answer within 120 s

Most calls answer in 15–25 s; each call waits up to 120 s by default. You get the token or ERROR_RECAPTCHA_TIMEOUT by then. Change the limit with timeout_s (5–270).

The token lasts 120 s

Google accepts a v2 token once, within 120 s of issue. expires_in_s says how many seconds are left, so ask for the token when the form is ready. A polled task read after that returns ERROR_TOKEN_EXPIRED.

Wrong keys fail fast

A mistyped sitekey, a v3 key, or a key not registered for the page's domain returns ERROR_RECAPTCHA_INVALID_SITEKEY or ERROR_RECAPTCHA_INVALID_DOMAIN right away instead of waiting for the time limit. Retrying will not help.

Task types

Every request carries a task.type.

task.typeCaptchaUse it when
ReCaptchaV3HSTaskProxyLessv3Recommended. Higher score, a few seconds slower: logins, sign-ups, checkouts.
ReCaptchaV3HSTaskv3Same, through your own proxy.
ReCaptchaV3TaskProxyLessv3Fastest, lower score. For sites that don't check the score strictly, and for testing.
ReCaptchaV3Taskv3Same, through your own proxy.
ReCaptchaV2TaskProxyLessv2Recommended for v2: “I'm not a robot” checkboxes, including image challenges.
ReCaptchaV2Taskv2Same, through your own proxy. Usually slower. Without proxy fields it runs as ProxyLess.
Getting a good v3 score

Send the exact pageAction and apiJs the page uses, and use a high-score type (ReCaptchaV3HSTask…). For the best scores on strict sites, add your own residential proxy. If you get ERROR_RATE_LIMIT, back off a few seconds and retry.

Fields

The request body is {"clientKey": "…", "task": {…}}. These go inside task.

BodyRequiredTypeWhat it is
clientKeyyesstringYour API key, in every request.
taskyesobjectThe task: the fields below.
FieldRequiredTypeDefaultWhat to send
typeyesstring—A task type from the table above.
websiteURLyesstring—Full URL of the page that shows the captcha, starting with https:// or http://. The token only works on this site.
websiteKeyyesstring—The site's reCAPTCHA key (6L…). v3: usually the render= value in the api.js URL; with render=explicit (or no render=), search the page source for the 6L… key. v2: the checkbox's data-sitekey.
pageActionv3string—The action name the page sends with reCAPTCHA, e.g. login, signup, checkout. It must match exactly: a token for the wrong action is rejected even with a good score. There is no default.
apiJsv3string—The reCAPTCHA script URL exactly as the page loads it, query string included, e.g. https://www.google.com/recaptcha/api.js?render=6L…. api.js or enterprise.js, on google.com or recaptcha.net; any render= key must equal websiteKey. Otherwise ERROR_TASK_INVALID.
timeout_snonumberv3 60 s, v2 120 sMost seconds to wait for a token. v3 default 60, v2 default 120; both from 5 to 270.
proxy fieldsnostrings, number—Only with a proxy task type (one without ProxyLess): see Advanced → Use your own proxy. Without them the task runs as ProxyLess.
isEnterprisenobooleanfrom apiJsSee Advanced → reCAPTCHA Enterprise.

Limits & timeouts

The values this service uses now.

LimitreCAPTCHA v3reCAPTCHA v2
New tasks per minute, per IP60
Default timeout_s60 s120 s
timeout_s range5–270 s
Token lifetime120 s, works once (see expires_in_s)
Polled results kept10 minutes
Poll getTaskResultevery 2 s
Your own proxyAllowed

Code in your language

Pick reCAPTCHA v3 or v2 and your language, then replace YOUR_API_KEY and the example values. Run this on your server or in a script: browsers block calls to this API from web pages, and your key would be visible.

curl -X POST "https://solver.mintbin.xyz/recaptcha/getToken" \
  -H "Content-Type: application/json" \
  -d '{
  "clientKey": "YOUR_API_KEY",
  "task": {
    "type": "ReCaptchaV3HSTaskProxyLess",
    "websiteURL": "https://example.com/my-account",
    "websiteKey": "6Lxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
    "pageAction": "login",
    "apiJs": "https://www.google.com/recaptcha/api.js?render=6Lxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
  }
}'

The reply

Replies are HTTP 200, failures included. Check errorId first.

FieldMeaning
errorId0 = success. Anything else = failure.
status"ready" when the token is in solution; "processing" while a polled task runs. Failures have no status.
solution.gRecaptchaResponseThe token.
taskIdThe task's id. Treat it as an opaque string.
elapsed_sSeconds the call took.
expires_in_sSeconds left before Google stops accepting the token (at most 120).
errorCode, errorDescriptionOn failure only: a stable code to branch on and a readable description.

Success

JSON
{
  "errorId": 0,
  "status": "ready",
  "solution": {
    "gRecaptchaResponse": "03AGdBq26...long-opaque-token..."
  },
  "taskId": "3f2a...",
  "elapsed_s": 8.3,
  "expires_in_s": 119
}

Failure

JSON
{
  "errorId": 12,
  "errorCode": "ERROR_CAPTCHA_UNSOLVABLE",
  "errorDescription": "No token this time. Retry; a second attempt often succeeds.",
  "taskId": "3f2a..."
}
Tokens are single-use and short-lived

A reCAPTCHA token works once, on the site it was issued for, within 120 s. Ask for it when the form is ready and submit before expires_in_s runs out.

Errors

Errors answer HTTP 200 too, so check errorId first: 0 is success (or status:"processing" while a task runs); anything else comes with an errorCode and an errorDescription. Branch on errorCode. An HTTP 502 or 504 means the service is briefly unavailable: retry in a few seconds.

Common codesMeaningWhat to do
ERROR_KEY_DOES_NOT_EXIST →The clientKey is missing or wrong.Send your key as clientKey in the JSON body.
ERROR_TASK_INVALID →A field is missing or invalid, for example websiteURL without http(s)://, a missing pageAction or apiJs (reCAPTCHA v3), or a non-numeric timeout_s.errorDescription names the field. Fix it and send the task again.
ERROR_RATE_LIMIT →Too many new tasks from your IP in the last minute.Back off for a few seconds, then retry.
ERROR_NO_SLOT_AVAILABLE →The service is busy or briefly unavailable.Retry in a few seconds.
ERROR_CAPTCHA_UNSOLVABLE →No token this time.Retry; a second attempt often succeeds.
ERROR_RECAPTCHA_TIMEOUT →No token within timeout_s.Retry, or send a larger timeout_s.
Advanced: polling, proxies and more

All endpoints

All paths are under /recaptcha. Send clientKey in the JSON body of every POST.

MethodPathWhat it does
POST/recaptcha/getTokenOne call: returns the token. Recommended.
POST/recaptcha/createTaskStarts a task and returns its taskId.
POST/recaptcha/getTaskResultSend {clientKey, taskId}. Returns status:"processing" until the token is ready, then "ready" with the solution. Results stay readable for 10 minutes.
POST/recaptcha/reportVerdictOptional feedback after the site has judged a token: send {clientKey, token, verdict: "accepted" | "rejected", websiteURL}. Later tokens for that site then come from the same place it accepted, and a place it rejected is not used for it again. Reply: {errorId: 0, applied: true | false}.
GET/recaptcha/healthIs the service up? Shows status, availability, and whether v3 and v2 are ready.
GET/recaptcha/whoamiThe IP address we see for you. Rate limits count per IP.

Polling instead of one call

Don't want to hold a connection open? POST the same body to /recaptcha/createTask. It returns a taskId right away. Then POST {clientKey, taskId} to /recaptcha/getTaskResult every 2 seconds until status is "ready". Stop as soon as errorId is not 0. Results stay readable for 10 minutes.

JSON
# createTask, then poll getTaskResult with the taskId
{ "errorId": 0, "taskId": "3f2a...", "status": "processing" }
{ "errorId": 0, "status": "processing" }
{ "errorId": 0, "status": "ready", "solution": { "gRecaptchaResponse": "03AGdBq26..." }, "taskId": "3f2a..." }
# 1) create the task
curl -X POST "https://solver.mintbin.xyz/recaptcha/createTask" \
  -H "Content-Type: application/json" \
  -d '{
  "clientKey": "YOUR_API_KEY",
  "task": {
    "type": "ReCaptchaV3HSTaskProxyLess",
    "websiteURL": "https://example.com/my-account",
    "websiteKey": "6Lxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
    "pageAction": "login",
    "apiJs": "https://www.google.com/recaptcha/api.js?render=6Lxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
  }
}'

# 2) poll every 2 seconds until "status": "ready"
#    (a non-zero errorId means the task failed: stop polling)
curl -X POST "https://solver.mintbin.xyz/recaptcha/getTaskResult" \
  -H "Content-Type: application/json" \
  -d '{ "clientKey": "YOUR_API_KEY", "taskId": "TASK_ID_FROM_STEP_1" }'

Use your own proxy

Pick a type without ProxyLess (ReCaptchaV3HSTask, ReCaptchaV3Task or ReCaptchaV2Task) and add to the task: proxyType (http, https or socks5), proxyAddress, proxyPort, proxyLogin, proxyPassword. The token is then issued from your proxy's IP. socks5 can't take a login: use http for proxies that need one. A call through your own proxy starts from a fresh IP and usually takes 20 to 40 seconds: give it a timeout_s of at least 60. If you get ERROR_NO_SLOT_AVAILABLE, retry after a few seconds with the same proxy. For reCAPTCHA v3, a residential proxy gives the best scores. Some keys may not send a proxy (ERROR_PROXY_NOT_ALLOWED).

cURL
curl -X POST "https://solver.mintbin.xyz/recaptcha/getToken" \
  -H "Content-Type: application/json" \
  -d '{
    "clientKey": "YOUR_API_KEY",
    "task": {
      "type": "ReCaptchaV3HSTask",
      "websiteURL": "https://example.com/my-account",
      "websiteKey": "6Lxxxx...",
      "pageAction": "login",
      "apiJs": "https://www.google.com/recaptcha/api.js?render=6Lxxxx...",
      "proxyType": "http",
      "proxyAddress": "192.0.2.10",
      "proxyPort": 8080,
      "proxyLogin": "user",
      "proxyPassword": "pass"
    }
  }'

reCAPTCHA Enterprise (isEnterprise)

Optional. We read it from apiJs: enterprise.js means reCAPTCHA Enterprise, with the same task types. If you send it and it disagrees with apiJs, you get ERROR_TASK_INVALID.

Next steps

May we load Google reCAPTCHA? It sets a cookie and sends Google details of your device and this visit, which helps us test our service. The site works the same either way, and you can change this anytime under Privacy choices at the bottom of each page. Privacy