Cloudflare Blocks Are a Routing Problem, Not a Retry Problem
Cloudflare automation failures need routing, not retries. Learn how to separate IP reputation, TLS, JavaScript, CAPTCHA, and rate-limit fixes in production.
Jul 29, 20268 min read
Command Palette
Search for a command to run...
#playwright
Articles tagged with #playwright
Treat Browser Fingerprints Like Production State
Browser fingerprint detection fails scrapers when profile state drifts. Learn concrete tactics for canvas, geo, TLS, sessions, and testing.
Jul 27, 20267 min readWire vs Skyvern: pick the failure domain before the automation tool
A practical engineering view of Wire vs Skyvern, comparing network-layer APIs, AI browser agents, failure modes, latency, and production tradeoffs.
Jul 6, 20268 min readCredential custody is the wrong abstraction for browser automation
Authenticated automation should treat passwords as just-in-time inputs and sessions as expiring leases, with clear custody and audit boundaries.
Jun 12, 20267 min readStop Choosing Scraping APIs by Feature Lists: Route by Failure Mode
A practical framework for Cloudflare scraping APIs: classify block signals, route proxy vs browser modes, and measure cost per verified extraction.
Jun 4, 20267 min read