Notes on scraping
the parts of the web
that don't want it.
Engine internals, anti-detection research, release notes. Written by people who ship the tool.
Beating Cloudflare Turnstile in 2026
Turnstile is now a JS-executed proof-of-work with device fingerprinting. Here's how the tier-2 engine clears it without a residential proxy.
TLS impersonation, and why your Python client is fingerprinted before byte one
The JA3 hash of a plain requests call is nothing like a real Chrome. curl_cffi is the fix, and this is what it actually does under the hood.
Scraping a marketplace end-to-end: from CLI call to Parquet
A walkthrough: point IntelliScrape at a real e-commerce site, crawl 12,000 SKUs, and end up with a queryable SQLite you can hand to a notebook.
IntelliScrape 0.9 — engine fallback, camoufox, hosted service beta
The tiered engine is the headline change. Also: canvas spoofing gets more accurate, and the hosted API opens in private beta.
"Human-like" is a measurable thing
Behavioral anti-bot is trained against telemetry. Mouse curves, dwell time, scroll acceleration — here's what to fake and what not to bother.