Home / Terms / Anti-bot detection
What is anti-bot detection?
Anti-bot detection is a website’s use of technical signals and challenges to identify or restrict automated visits. Scraping vendors respond with protected, stealth, premium-routing, browser, CAPTCHA, or named-bypass modes, but those labels do not represent one standard capability or billing rule. (ScraperAPI credit rules, ScrapeOps request costs, and ZenRows pricing rules, checked 2026-07-26.)
Anti-bot detection across the APIs we track
| API | Mode | Published rate | Compared with a plain request | Checked |
|---|---|---|---|---|
| Browse AI | Protected-site rate | 2 credits per task | 2 × a plain request | Oct 9, 2026 |
| Decodo Web Scraping API | Premium proxies | $1.00 per 1,000 requests | 2 × a plain request | Oct 6, 2026 |
| HasData | Residential proxies | 5 credits per request | 5 × a plain request | Oct 8, 2026 |
| Scrape.do | Residential proxies | 10 credits per request | 10 × a plain request | Oct 9, 2026 |
| ScrapeGraphAI | Stealth mode | 5 credits per request | not comparable to its plain rate | Oct 9, 2026 |
| ScrapeOwl | Premium proxies | 25 credits per request | 25 × a plain request | Oct 9, 2026 |
| ScrapeOwl | Residential proxies | 10 credits per request | 10 × a plain request | Oct 9, 2026 |
| ScraperAPI | Premium proxies | 10 credits per request | 10 × a plain request | Oct 9, 2026 |
| ScraperAPI | Protected-site rate | 10 credits per request | 10 × a plain request | Oct 9, 2026 |
| ScraperAPI | Stealth mode | 30 credits per request | 30 × a plain request | Oct 9, 2026 |
| ScrapingAnt | Residential proxies | 25 credits per request | 25 × a plain request | Oct 9, 2026 |
| ScrapingBee | Stealth mode | 75 credits per request | 75 × a plain request | Oct 7, 2026 |
| ScrapingBot | Stealth mode | 75 credits per request | 75 × a plain request | Oct 9, 2026 |
| SOAX Web Data API | CAPTCHA solving | 1 request, unit not published | not comparable to its plain rate | Sep 2, 2026 |
| SOAX Web Data API | Residential proxies | 1 request, unit not published | not comparable to its plain rate | Sep 2, 2026 |
| WebScraping.AI | Residential proxies | 10 credits per request | 10 × a plain request | Oct 9, 2026 |
| WebScraping.AI | Stealth mode | 50 credits per request | 50 × a plain request | Oct 9, 2026 |
| ZenRows | Premium proxies | 10 credits per request | 10 × a plain request | Oct 9, 2026 |
Vendors scope these modes differently: one mode may cover a site another vendor prices as its hardest tier.
In more detail
The vendor examples below were read on the dates given in the text; the table above holds each vendor's current figure.
Anti-bot handling matters to a bill because a protected request can consume many more credits than the basic route. Some vendors require the buyer to select the mode, while others can escalate routing automatically. A forecast based only on ordinary pages can therefore undercount credits when the production target uses a protection-specific configuration. These are billing rules, not outcome measurements. (ScraperAPI credit rules, Scrapfly billing rules, and ZenRows pricing rules, checked 2026-07-26.)
ScraperAPI publishes ten credits per scrape for named Cloudflare, Turnstile, DataDome, and PerimeterX bypass modes. Its stealth mode costs thirty credits, while ultra-premium routing with rendering has a larger combined cost. The documentation does not define how every bypass, target, and parameter cost combines, so those interactions remain unknown. (ScraperAPI credit rules, checked 2026-07-26.)
ScrapeOps publishes three generic bypass levels at ten, twenty-five, and forty credits. It separately lists PerimeterX at twenty-five and DataDome at forty, while premium levels and JavaScript combinations use other totals. These are replacement or combination rules, not one add-on that can be applied blindly to every base request. (ScrapeOps request costs, checked 2026-07-26.)
ZenRows publishes a twenty-five-credit protected mode alongside one-credit basic, five-credit JavaScript, and ten-credit premium-proxy modes. The four Universal Scraper modes replace one another. Failed requests and internal retries are free, while HTTP 404 and 410 responses are billed as successful. (ZenRows pricing rules, checked 2026-07-26.)
Scrapfly’s Automatic Scraping Protection can dynamically upgrade the proxy pool, changing the base request cost. JavaScript and excess response bandwidth are additive, so an automatically upgraded request can also accumulate those meters. Its failure-protection billing has conditions and can be disabled when eligible failures exceed the published traffic threshold. (Scrapfly billing rules, checked 2026-07-26.)
Before signing up, identify the target’s actual challenge and match it to a documented mode. Check whether escalation is manual or automatic, how named bypasses combine with rendering and proxies, which statuses are billed, and whether the response exposes the chosen route and debit. Set a maximum acceptable per-request cost when the vendor permits it, and test each target class separately. (ScraperAPI credit rules, ScrapeOps request costs, ZenRows pricing rules, and Scrapfly billing rules, checked 2026-07-26.)