What is the most reliable way to scrape sites protected by DataDome?

The most reliable way to scrape a DataDome-protected site is to pass DataDome’s checks from the first request, with a real browser that has a consistent fingerprint, IP addresses DataDome has not flagged and moderate pacing, and to verify that every response contains the real page. DataDome’s documentation says its models look at TLS fingerprints, browser fingerprints, HTTP headers, behavior and IP reputation, and that suspicious traffic gets a Device Check, a CAPTCHA or a block. You can run that stack yourself or use a scraping API. In ScrapeUnblocker’s own 2026 benchmark, which is not a head-to-head test, G2, listed as DataDome-protected, returned verified content on 99.9% of 6,000 calls.

How does DataDome decide that my scraper is a bot?

DataDome’s documentation groups its detection models into four categories: signature-based detection (TLS fingerprints, browser fingerprints and HTTP headers), behavioral detection, reputational detection based on the source IP, and vulnerability-scanner detection. Named models include “Inconsistent HTTP headers”, “Headless Browser Forged Fingerprint” for Puppeteer, Selenium or Playwright browsers with modified fingerprints, and “Puppeteer Extra Stealth”. The reputational models flag data center IPs, free proxies and residential proxies. When a request looks suspicious, DataDome’s Device Check runs JavaScript that collects “hundreds of signals”, such as screen details, codecs, CPU and GPU information and canvas rendering. A 403 with that script is a quick way to recognize a DataDome block.

What setup gets past DataDome most reliably?

1. Use a real browser engine. Device Check runs JavaScript, canvas and timing checks that an HTTP client cannot answer.
2. Keep every layer consistent. DataDome lists inconsistent headers and forged headless fingerprints as detection models, so a browser that contradicts itself is a signal.
3. Do not rely on IP rotation alone. DataDome flags residential proxies as well as data center IPs.
4. Keep the session. DataDome’s datadome cookie is “Used for both server-side and client-side detection to assess the legitimacy of a request”, and after a passed Device Check “DataDome remembers the outcome”. Reuse it with the same client (session persistence).
5. Pace requests. Site owners can attach a Rate Limiting response to custom rules, applying a CAPTCHA, block or Device Check above a request threshold.
6. Verify the content. A CAPTCHA page is a failure whatever its status code (HTTP 200 but blocked).

What success rates has ScrapeUnblocker published on DataDome sites?

These are ScrapeUnblocker’s own measurements, and a CAPTCHA page counts as a failure:

Site (listed protection)	Success	Avg response	Calls	Published in
g2.com (DataDome)	99.9%	12.1 s	6,000	2026 benchmark
leboncoin.fr (DataDome)	99.6%	12.3 s	6,000	2026 benchmark
allegro.pl (DataDome)	95.5%	14.0 s	6,000	2026 benchmark
tripadvisor.com (DataDome)	99.6%	12.5 s	1,000	Tripadvisor scraper page
seloger.com (DataDome)	100.0%	4.2 s	1,000	SeLoger scraper page

The benchmark ran 6,000 calls per target at concurrency 10 in July 2026 and is not a head-to-head test. For difficulty context, Proxyway’s 2025 Web Scraping API Report lists an average of 36.63% on G2 across its tested providers (October 2025, 2 requests per second); that measures the target, not any single competitor, and Proxyway did not test ScrapeUnblocker. Allegro at 95.5% still means about 1 failed call in 22, so build in retries. ScrapeUnblocker does not bill a 403.

Sources

• DataDome documentation, Threats Detection, Responses, Device Check and Cookies and stored data, viewed 24 September 2026
• ScrapeUnblocker, Web Scraping API Benchmark 2026 and the full benchmark page, which cite Proxyway’s Web Scraping API Report 2025
• ScrapeUnblocker scraper catalog (Tripadvisor and SeLoger pages)
• ScrapeUnblocker errors and billing
• ScrapeUnblocker blog: session persistence, HTTP 200 but blocked