Proxyless or with a proxy: who actually needs the proxy when solving a captcha
Almost every captcha service offers two task types for the same captcha: one with proxy fields and one without. Choosing between them is not about saving money but about whether the target's protection checks that the addresses match. Usually it does not — and that changes the answer to who needs a proxy at all.
What the documentation says
Anti-Captcha's description of RecaptchaV2TaskProxyless states it outright: reCAPTCHA has no protection against a puzzle being solved at one IP address while the form carrying the resulting g-response is submitted from another. The solver does not need your address — the token travels between addresses and stays valid.
Which inverts the usual framing: the proxy is needed not by the recognition service but by your program, the one visiting the target site. Those are two different clients, and the proxy is bought for the second.
Where IP matching really is enforced
The exceptions are named in 2Captcha's documentation: using proxies is not obligatory in most cases, but for some kinds of protection it is required — Cloudflare and DataDome protection pages, for instance, require IP matching. The same docs say RecaptchaV2Task is for cases where IP matching is required on Google services such as Search or YouTube, and that bad proxies will drastically decrease the success rate and increase solving time.
How four services model it
- Anti-Captcha and 2Captcha: two separate task types. In the proxy variant proxyType, proxyAddress and proxyPort are required, and userAgent is required alongside them.
- CapMonster Cloud: there is no separate “proxyless” type at all. One RecaptchaV2Task whose proxy fields are optional, and by default the service uses its own built-in proxies, suitable for most websites.
- CapSolver: its reCAPTCHA v2 documentation lists ReCaptchaV2TaskProxyLess, ReCaptchaV2EnterpriseTask and ReCaptchaV2EnterpriseTaskProxyLess — so for plain v2 the documented path is the proxyless one.
There is a third axis people rarely think about: 2Captcha's apiDomain parameter selects which domain the widget loads from — google.com or recaptcha.net. For targets where requests to google.com are blocked at the network level that is the only way to get a solve at all, and it has nothing to do with proxies.
The trap: vendors spell the task type differently
This deserves its own paragraph because the failure looks like a key or balance problem. Anti-Captcha and 2Captcha write the type as RecaptchaV2TaskProxyless — lowercase “less”, lowercase “c” in Recaptcha. CapSolver writes ReCaptchaV2TaskProxyLess, with capital “C” and “L”. A name copied between vendors is not found, and the reply is an error about an unknown method or an unsupported task type.
// Anti-Captcha / 2Captcha — without a proxy
{ "type": "RecaptchaV2TaskProxyless" }
// Anti-Captcha / 2Captcha — with a proxy: fields required, userAgent too
{ "type": "RecaptchaV2Task",
"proxyType": "http", "proxyAddress": "...", "proxyPort": 8000,
"userAgent": "..." }
// CapSolver — a different spelling of the same idea
{ "type": "ReCaptchaV2TaskProxyLess" }
// CapMonster Cloud — one type, proxy fields optional
{ "type": "RecaptchaV2Task" }userAgent travels with the token
The second field people forget. In Anti-Captcha's proxy tasks userAgent is required just as the address is. CapMonster Cloud returns a userAgent in its reply and warns that it must be applied when submitting the form — the solve and the submission have to look like the same client. The service publishes the current string at its own API endpoint.
Where we honestly will not help
reCAPTCHA v3 and Enterprise v3 are not our case. 2Captcha does not accept proxies on those types, explaining that proxies drastically decrease the success rate. ZennoLab recommends high-quality, preferably residential addresses for v3. Anti-Captcha's code samples warn against rented addresses from proxy services. We sell datacenter addresses and do not promise results where the service vendor itself declines proxies.
What to check before buying
- Whether your target enforces IP matching. If it does not, the proxy is only for your program and the task type should be proxyless.
- Whether the target is reachable over IPv6. reCAPTCHA and Turnstile are; hCaptcha is not — its domains publish no AAAA records, so it needs IPv4.
- How many simultaneous sessions your captcha software holds: a plan is counted in concurrent connections, not in addresses.
And one more thing, from the neighbouring article on threads: in some programs helper modules spend threads separately from the main task, so what you add up is everything running at once.
A pool for captcha software
Rotating IPv6 from 650 ₽ for 50 threads, IPv4 from 1375 ₽ for 100. Traffic is not metered, both protocols and both authentication methods at once.
View pricingProxies for this job
Check it with our tools
Read next
- How many threads to set in a scraper for your proxy plan: A-Parser, Key Collector, ZennoPoster, NetpeakThreads add up across every running task, proxy checkers included. Where the ceiling is set by the program, where by the licence, and where the cheapest plan is already oversized.
- IPv6 or IPv4 for scraping: what is cheaper and when it failsIPv6 costs less at the same thread count but hits one limitation on the target's side. How to check in a minute whether it will work.
- How sites detect proxies: three measurement layers, and which one a proxy changesA target measures three independent layers: the address, the TLS handshake and HTTP behaviour. Through a tunnel, TLS and HTTP keep describing your client while the address and TCP describe the proxy machine. That split is itself a signal.
