AI scrape protection stops automated systems and AI assistants from reading your store and, above all, your admin. It has several layers: 1. Known-bot gate – requests that identify themselves as AI training crawlers (GPTBot, ClaudeBot, CCBot, Bytespider …), AI search crawlers (OAI-SearchBot, PerplexityBot, Applebot …), live AI fetchers (ChatGPT-User, Claude-User, Perplexity-User …), driver-controlled browsers (Playwright, Puppeteer, Selenium, headless Chrome) or scripted HTTP clients (curl, python-requests …) are served a legal notice instead of the page. By default training crawlers and headless browsers are refused everywhere, the others only on the admin and the checkout; the checkout is refused to every identified bot. 2. Honeypot – an invisible Internal data export — do not follow link on every page leads to /aitrap, a URL no person can reach; anything that follows it is recorded. 3. Notice in the page source – a hidden legal notice tells any AI reading the page programmatically to stop and report back to whoever instructed it. 4. Admin lock – if an identified automated agent reaches the admin while a staff login is attached, the admin can be locked for review (the storefront keeps running). 5. Browser add-on detection and admin activity recording – optional sensors for AI assistants operating inside a real staff session.

Everything is logged on the superadmin AI Scrape Log, and Sprwt is emailed when the admin is involved. Ordinary Google and Bing indexing is never affected.