# AIO Scanner · Free AI-readability check · Onto > Drop any URL. Get a full AI-readiness report in seconds — React Tax, semantic richness, hallucination risk, content negotiation. Free. No signup. The same scanner we used to grade Stripe, Linear, and Resend. **Source:** /scanner **Extracted:** 2026-05-20T20:59:17.346Z --- Free toolAIO Scanner ## What do AI agents _actually see on your site?_ Drop any public URL. Get a full AI-readiness report in seconds — React Tax, semantic richness, hallucination risk, content negotiation status, and the exact bytes a Claude or GPT agent would receive. **Free. No signup. No card.** Run AIO scan 126 Sites scanned 51/100 Average score ~4s Median scan time 01 // What the scanner checks ### Three weighted pillars. Five live checks. The AIO score is subtractive — every penalty traces to a specific structural issue. Same model we used to publish the [Stripe, Linear, and Resend teardowns](/case-studies). Pillar01 React Tax What fraction of the bytes shipped to an agent is visual scaffolding it can't use — CSS rules, JS bundles, layout divs. Most SaaS sites score 60–95% noise. Pillar02 Semantic Richness Whether the HTML tells a structured story: heading hierarchy, JSON-LD schema, landmark elements. The scaffold an LLM needs to verify facts. Pillar03 Content Negotiation Whether the server responds differently to Accept: text/markdown or known AI User-Agents. The single biggest agent-readiness signal — and almost nobody ships it. Check04 Robots / WAF posture Whether GPTBot, ClaudeBot, PerplexityBot are explicitly allowed, and whether your WAF lets them through without challenge pages. Check05 Heading hierarchy Whether H1–H3 actually map to your content structure, or whether they're decorative — the diff between a parseable page and a guessable one. 02 // What the report looks like ### Three live examples from sites you know. Before scanning your own URL, see what the report actually contains. We've published full teardowns of three well-known SaaS pricing pages — the kind of analysis your scan will produce. [ 20/100 AI-hostile Linear pricing 1.8 MB of JavaScript for a 0.3 KB page ](/case-studies/linear-pricing)[ 65/100 Needs work Stripe pricing 690 KB of noise for a 0.8 KB story ](/case-studies/stripe-pricing)[ 90/100 Agent-ready Resend pricing Already serves AI-ready Markdown ](/case-studies/resend-pricing) 03 // Before you scan ### Common questions. 01Is the scanner really free? Yes. No signup, no card, no usage limit on a single URL scan. We rate-limit per IP at a generous threshold to keep the service up — that's it. 02How accurate is the AIO score? The score is subtractive — every penalty traces to a specific structural issue (missing headings, content negotiation absent, schema unparseable, etc.). We expose the full breakdown so you can see why a page scored what it scored, not just the number. 03Does the scanner actually fetch my page like a real AI agent? Yes. We send Accept: text/markdown and known AI User-Agents (GPTBot, ClaudeBot, PerplexityBot) and compare what comes back against the human-browser response. Same wire-level behaviour as a real crawler. 04Can I share the result? Yes. After a scan completes, you'll see a permalink URL with your scanned domain encoded — copy it from the address bar to share the full report. Public URLs are scannable by anyone. 05What's the difference between this and Onto's paid Read API? The scanner gives you a one-shot report — diagnostic. The [Read API](/read-api) gives you the same scoring on every request, programmatically, plus the clean Markdown output. Use the scanner to decide if you have a problem; use the API to actually fix it. 06Do you store the URLs I scan? We log the scanned URL and resulting score for our public 'sites scanned' counter, but we never store the page contents beyond the 10-minute response cache. No personal data leaves your browser. Ready when you are ### Scan your site. See what an agent sees. Scroll back up, paste your URL, hit scan. Report renders in roughly four seconds. Back to scanner// Free · No card · 4 seconds