A CapSkip result is readable once, so a Celery retry that resumes from a stored captcha id gets nothing. Here is the task shape that survives a worker crash, the two time limits to set, and the broker option that quietly doubles your solves.

hrequests spoofs the TLS fingerprint. It does not read CAPTCHAs, and CapSkip does not spoof fingerprints, so the two solve different halves of the same block. Here is how to fetch the sitekey, solve it, and post the token back on the session that earned the cookies.

Activepieces Cloud runs Code steps in a V8 isolate that silently discards the npm packages you add, so the CapSkip Node SDK never arrives. The HTTP piece works in every sandbox mode. Here is the flow, and the network setting that blocks a local solver.

Windmill resolves PyPI imports for you, so the CapSkip SDK needs no install step. What does need thought is which machine the worker runs on, because that decides whether the solver stays on the loopback address or moves to Server mode.

Pipedream code steps run real Node.js, so the CapSkip SDK works unchanged. The catch is the execution timeout, which is shorter than a reCAPTCHA solve. Here is the one-step version, the flow.rerun version that survives a slow solve, and the Python equivalent.

Scrapling ships solve_cloudflare, which handles Cloudflare and nothing else. reCAPTCHA, GeeTest and image CAPTCHAs go in through page_action instead. Here is the injection, the isolated execution context that silently breaks it, and a full working script.

Node-RED can call a CAPTCHA solver two ways: http request nodes against the raw API, or the Node SDK inside a function node. Here is both, plus the settings.js change that makes the second one possible and the reason a Raspberry Pi install needs Server mode.

Temporal workflow code has to be deterministic, so a CAPTCHA solve can only live in an Activity. Here is the activity, the retry policy that skips the failures that will never pass, and why a durable workflow makes token expiry easier to get wrong than anywhere else.

Four CapSkip error codes come back before the solver has looked at a CAPTCHA at all. They are about the shape of your HTTP request, and each one names a different mistake. Here is what each means and how to fix it.

An httpx captcha flow has no browser, so nothing fills the token field for you. Read the sitekey out of the HTML, solve it locally, then post the token as an ordinary form field on the same client.