How do I set up stealth scraping?

Stealth scraping collects web data without triggering anti-bot defenses. The stack: headless browser (Playwright recommended — better stealth than Puppeteer), anti-detect patches (stealth plugin, realistic fingerprints), residential proxies (rotate per session, geo-matched to target), and rate limiting (2-5 second delays between requests). Architecture: 1) Use Playwright with stealth plugin to eliminate headless detection signals. 2) Route through residential proxies with sticky sessions (same IP per browsing session). 3) Implement realistic behavior: random delays, mouse movement simulation, scroll before extracting. 4) Handle CAPTCHAs with solving services as fallback. 5) Rotate user agents matching the proxy location. 6) Monitor success rates — if block rate exceeds 10%, adjust strategy. For high-volume scraping (10,000+ pages/day): distribute across multiple machines, use session-based proxy pools, implement retry logic with exponential backoff, and cache results to avoid re-scraping. Legal note: scraping publicly available data is generally legal (hiQ v. LinkedIn), but violating Terms of Service and accessing private data behind authentication raises legal questions.