nonecap

Official Python client for the NoneCap hCaptcha solver API.

Submit a captcha, get back a token. The client handles the polling, the timeouts, and the error cases so you don't write the request loop yourself. Sync and async, fully typed.

Install

pip install nonecap

Python 3.9+. The only dependency is httpx.

Quick start

Grab an API key from dashboard.nonecap.com, then:

from nonecap import NoneCap

nc = NoneCap(api_key="nc_live_...")

solve = nc.solve(
    type="hcaptcha",
    sitekey="10000000-ffff-ffff-ffff-000000000001",
    url="https://example.com/login",
)

print(solve.token)  # the hCaptcha token, ready to submit
print(solve.resp_key)  # hcaptcha.getRespKey() equivalent, for sites that verify the pair
print(solve.user_agent)  # submit the token with this User-Agent: some sites check the browser version
print(solve.warnings)  # e.g. proxy_session_refused: tokens you reported from this proxy session were refused; a new one is more likely to pass

solve() submits the captcha and waits until it's done, using the API's long-poll so you aren't hammering it with requests. It returns the solved solve, or raises if the solve fails or your timeout runs out.

Async

Same surface, awaited. Use it as an async context manager so the connection pool gets cleaned up:

import asyncio
from nonecap import AsyncNoneCap

async def main() -> None:
    async with AsyncNoneCap(api_key="nc_live_...") as nc:
        solve = await nc.solve(type="hcaptcha", sitekey="...", url="https://example.com")
        print(solve.token)

asyncio.run(main())

Cancelling a solve

solve() is the simple path when you just want a token. When you need to hold a reference you can cancel — for clean shutdown, freeing a worker slot, or stopping early — use solves.start(). It submits the solve and hands back a SolveHandle right away, with the id already populated, instead of blocking on the result:

handle = nc.solves.start(type="hcaptcha", sitekey=sitekey, url=url)
print(handle.id)  # available immediately

# Wait for it, just like solve():
solve = handle.result(timeout=120)
print(solve.token)

Holding the handle lets you stop a solve you no longer need — say, on shutdown or when a parallel attempt already won — instead of waiting on its result:

handle = nc.solves.start(type="hcaptcha", sitekey=sitekey, url=url)

# ... elsewhere / later, to stop it:
handle.cancel()

handle.result() long-polls until the solve finishes and returns it, raising SolveFailedError / SolveTimeoutError like solve() does. The terminal outcome is memoized, so once the solve has settled, calling result() again replays it for free; a SolveTimeoutError is not memoized, so you can call result() again with a larger timeout to keep waiting. handle.cancel() stops a pending or in-flight solve and returns its final state — if the solve already finished, that's not an error, you just get the completed solve back.

The async client mirrors this — await the start, the result, and the cancel:

async with AsyncNoneCap(api_key="nc_live_...") as nc:
    handle = await nc.solves.start(type="hcaptcha", sitekey=sitekey, url=url)
    print(handle.id)

    # Either wait for the outcome ...
    solve = await handle.result(timeout=120)
    print(solve.token)

    # ... or, on another handle, cancel it instead of awaiting:
    other = await nc.solves.start(type="hcaptcha", sitekey=sitekey, url=url)
    await other.cancel()

Cancelled solves are never charged, and a solve you simply abandon expires uncharged at the server deadline — nothing is billed unless a solve actually succeeds. So cancel() is for cleanup and early-stop, not cost protection.

Cleaning up after a solve() timeout

solve() blocks until the solve settles, so it gives you no handle to cancel a solve while it's still running — for that, use solves.start() above. The one thing the solve() path offers is cleanup after a timeout: when solve() raises SolveTimeoutError, the error usually carries the in-flight solve_id (and last-known solve), so you can cancel the solve the wait gave up on. Guard on solve_id being present — if the very first submission times out at the transport level, no id has been assigned yet, so solve_id is None.

from nonecap import SolveTimeoutError

try:
    nc.solve(type="hcaptcha", sitekey=sitekey, url=url)
except SolveTimeoutError as err:
    if err.solve_id is not None:
        nc.solves.cancel(err.solve_id)
    else:
        raise

Handling failures

Every error this library raises extends NoneCapError, so you can catch the whole family or pick out the one you care about.

from nonecap import (
    NoneCap,
    SolveFailedError,
    InsufficientCreditsError,
    RateLimitError,
)

try:
    solve = nc.solve(type="hcaptcha", sitekey=sitekey, url=url)
except SolveFailedError as err:
    # The message says what happened, what to do, and that nothing was charged.
    print(err.solve.error.message if err.solve.error else err)
    if err.retryable:
        schedule_retry()
except InsufficientCreditsError:
    print("Out of credits. Top up at dashboard.nonecap.com")
except RateLimitError as err:
    time.sleep(err.retry_after or 1)

The subclasses are AuthenticationError (401), PermissionDeniedError (403), InsufficientCreditsError (402, with KeyCreditLimitError when one API key hit its own cap), ValidationError (422/400, with a param naming the bad field; PayloadTooLargeError for 413 and UnsupportedMediaTypeError for 415), NotFoundError (404), ConflictError (409), RateLimitError (429, with ConcurrencyLimitError, SitekeyRateLimitedError, ProxyUnavailableError (your own proxy is refusing connections) and RateCappedError (your account reached its submit rate cap on this sitekey, or your own-proxy sessions for it are at their limit) telling them apart; retry_after is the seconds the API asked you to wait), APIError (5xx, with ServiceUnavailableError for a maintenance pause), APIConnectionError and APITimeoutError (the request never landed), and SolveTimeoutError (your solve() budget ran out). Every error from a response carries request_id, the id to quote to support.

SolveFailedError carries the full solve. solve.error.code is a SolveErrorCode, solve.error.reason a typed sub-reason or None (proxy_rejected, sitekey_rate_limited, …), and solve.error.retryable says whether resubmitting the same request unchanged can succeed; err.retryable, err.reason and err.solve_code are shortcuts to those fields. Failed solves are never charged.

Enterprise captchas

For hcaptcha_enterprise, rqdata is optional: many enterprise sites never issue one. When the page does pass an rqdata blob to its widget, send it, fresh for each challenge:

solve = nc.solve(
    type="hcaptcha_enterprise",
    sitekey=sitekey,
    url=url,
    rqdata="...",  # only when the site provides one
)

Proxies

Pass a proxy as a dict or a URL string. The solve runs through it, and the bytes are metered back on the solve.

nc.solve(
    type="hcaptcha",
    sitekey=sitekey,
    url=url,
    proxy={"scheme": "http", "host": "1.2.3.4", "port": 8080, "username": "u", "password": "p"},
    # or: proxy="http://u:p@1.2.3.4:8080"
    # scheme can be http, https, socks5, socks5h, or socks4 (default http)
    # e.g. proxy="socks5://u:p@1.2.3.4:1080"
)

Reporting token acceptance

A token that hCaptcha blessed can still be refused by the site you send it to. Tell us what happened and we tune minting against your real acceptance rate — it's free, and it's the fastest way to get a regression on your sitekey noticed.

Keep the solve_id next to the token you submit downstream, then report the verdict:

solve = nc.solve(type="hcaptcha", sitekey=sitekey, url=url)
ok = submit_to_the_site_you_are_automating(solve.token)

nc.feedback.report(
    solve.id,
    outcome="accepted" if ok else "rejected",
    reason=None if ok else "session invalidated",  # optional, the code your target returned
    context=None if ok else "3rd retry, rotating residential pool",  # optional, anything you think would help us diagnose it
)

At volume, buffer the verdicts and flush them in one call. Reports over 500 are split across requests for you:

batch = nc.feedback.report_many([
    {"solve_id": "solve_01J...", "outcome": "accepted"},
    {"solve_id": "solve_01J...", "outcome": "rejected", "reason": "...", "context": "..."},
])

# Items resolve independently, so the call succeeds even when some are
# rejected — check `failed` rather than relying on a raised error.
if batch.failed:
    print([r for r in batch.results if r.status == "error"])

outcome is one of accepted, rejected, unknown (submitted, verdict unclear), unused (never submitted), or error (downstream broke for a non-token reason). Only accepted and rejected count toward the acceptance rate.

Reporting the same solve again corrects the earlier verdict, so retries and late fixes are safe. You can report any of your own solved solves within ~30 days of the solve; corrections to something you already reported are never cut off by that window. On AsyncNoneCap both methods are coroutines.

Lower-level API

solve() is the convenient path. When you want control over submission and polling, the resource methods map one to one to the REST API:

# Submit without waiting: returns immediately with a pending solve
pending = nc.solves.create(type="hcaptcha", sitekey=sitekey, url=url)

# Submit and hold the connection up to 30s for it to finish
maybe_done = nc.solves.create(type="hcaptcha", sitekey=sitekey, url=url, wait=30)

# Poll one solve, long-polling up to 30s
solve = nc.solves.retrieve(pending.id, wait=30)

# Cancel a pending or in-flight solve
nc.solves.cancel(pending.id)

# List a page of solves
page = nc.solves.list(limit=50, status="solved")

# Or iterate every solve, newest first
for s in nc.solves.list_all():
    print(s.id, s.status)

# Your account and credit balance
me = nc.me()
print(me.credits_balance)

On AsyncNoneCap the same methods are coroutines, and list_all() is an async iterator (async for s in nc.solves.list_all()).

Configuration

NoneCap(
    api_key="nc_live_...",              # required
    base_url="https://api.nonecap.com", # override if you need to
    timeout=100.0,                      # per HTTP request, seconds
    http_client=my_httpx_client,        # inject your own httpx.Client
)

solve() takes its own timeout (seconds, default 180) for the overall wait.

Typing

The package ships a py.typed marker and full inline annotations. Solves come back as frozen dataclasses with the exact field names the API uses (solve.token, solve.credits_charged, solve.queue_ms), so what you read in the API reference is what you get in code.

License

MIT, see LICENSE.