# AgentBI > 225 storefronts scanned once a day from KR; every measurement is published as raw JSON so anyone can recompute a row. Today (2026-11-23): 126/225 reachable, 63 serving a UCP profile. ## Start here - [Home](https://agentbi.tech/): paste a URL, get an agent-readiness report (10 checks, ~15s) - [Leaderboard](https://agentbi.tech/leaderboard): today's scores for the whole pool, filterable by platform and score band - [Raw daily batch](https://agentbi.tech/batch.json): every field we measured today, no key required ## Data - [Protocols](https://agentbi.tech/protocols): UCP / ACP / AP2 deployment counts - [Crawlers](https://agentbi.tech/crawlers): which AI agents each site blocks in robots.txt - [Trend](https://agentbi.tech/trend): day-over-day movement of the same sites - [Categories](https://agentbi.tech/categories): the same data grouped by industry ## About - [About, crawling behaviour, opt-out](https://agentbi.tech/about): honest user agent, rate limits, how to be removed - [CLI on PyPI](https://pypi.org/project/agentbi/): uvx agentbi — same checks from your terminal ## For agents - /batch.json is the machine-readable source behind every number on this site (JSON, Cache-Control: max-age=1800). - /api/check runs a live scan and is rate-limited; /robots.txt disallows it. Use /batch.json instead. - Scores are heuristic, not an audit. See /about.