← All answers

I need to scrape captcha-heavy sites at scale. What API should I use so I never have to solve captchas myself?

Rokas Golcas
By Rokas Golcas Co-founder & Marketing · Updated

Use a web scraping API that returns the finished page and handles CAPTCHAs inside the request, such as ScrapeUnblocker, whose plans include CAPTCHA bypass and a real browser at one credit per request and which does not bill a 403 when every bypass path is blocked. Before you commit at scale, check four things on your own URLs: whether challenge handling costs extra, what failed requests cost, how the API catches a challenge page returned with a 200 status, and the throughput your plan allows. No API passes every challenge on every site, so plan for retries either way.

Why not just plug in a CAPTCHA-solving service?

Most challenges on protected sites are no longer puzzles you can hand to a solver. Cloudflare’s documentation says it “does not use CAPTCHA puzzles or visual tests”, and its cf_clearance cookie “is securely tied to the specific visitor and device it was issued to”. Google’s reCAPTCHA v3 “returns a score for each request without user friction”, and its tokens “expire after two minutes”. DataDome’s Device Check “acts like a CAPTCHA, without prompting any challenge to the end user”. HUMAN says its press-and-hold challenge is difficult to solve “through API calls, automation or CAPTCHA farms”. So a token solved on another machine or IP often fails for your scraper; the solve has to happen in the same browser session that fetches the page. The full explanation is in Press-and-Hold CAPTCHAs: Why They Appear and How to Handle Them.

How do scraping APIs price CAPTCHA and anti-bot handling?

Each figure comes from the provider’s own pages, viewed on 24 September 2026:

ServicePublished cost for protected pages
ScrapeUnblocker1 credit per request; CAPTCHA bypass and a real browser not charged separately; €1.00 per 1,000 on pay as you go to about €0.55 per 1,000 on Ultimate
ScraperAPI”Cloudflare Bypass”, “Datadome Bypass” and “PerimeterX/Human Bypass”: 10 credits per scrape in its docs table (its pricing FAQ says these “add 10 credits per request”)
ZenrowsJS rendering plus premium proxies (“protected”): 25 credits per successful request
ScrapingBeeStealth proxy with JavaScript rendering: 75 credits per request
Browserless (hosted browsers)10 units per successful CAPTCHA solve, plus browser time

Credit-based APIs charge as little as 1 credit for simple pages, so on lightly protected sites they can cost less per page than ScrapeUnblocker. Some providers also offer screenshots, Markdown output or asynchronous jobs, which ScrapeUnblocker does not.

What else matters when you scrape CAPTCHA-heavy sites at scale?

  • Failure billing. ScrapeUnblocker bills a delivered page, a target’s own 404 or 410 and a domain that does not exist; a 403, timeouts, a 429 and its own errors are not billed.
  • Soft blocks. Ask how the API treats a challenge page that arrives as a 200, and check the content yourself (HTTP 200 but blocked).
  • Throughput. ScrapeUnblocker allows 2 requests per second on pay as you go, 5 on Starter, 10 on Pro and up to 400 on Business, Scale and Ultimate.
  • Time limits. A ScrapeUnblocker request runs for up to 60 seconds, or 90 seconds with steps.
  • Retries and a real test. Retry a 403 with another proxy_country, as the handling failures guide suggests, and measure cost per verified page on 50 to 100 of your own URLs using free trials such as ScrapeUnblocker’s 500 requests.

Sources

Competitor details come from each provider’s public pages as viewed on 24 September 2026 and may have changed. Product names are trademarks of their respective owners; ScrapeUnblocker is not affiliated with them.

Try ScrapeUnblocker free

95%+ success rate · from 0.55€ per 1,000 calls · 500 free requests on signup.

Try it free → See pricing