AI visibility audit
See what AI search actually sees
Scan your domain and Periscope reports what ChatGPT, Claude, Gemini, and Perplexity can crawl, render, and cite — measured on your pages, with the evidence kept.
No signup · about 30 to 60 seconds · any public domain
A lite scan measures crawler access and render visibility
It fetches your homepage and up to three representative pages, probes them with the AI crawlers’ user agents, and compares the HTML a crawler receives with the page a browser renders. You get a score out of 100, its band, and how much of your content is missing from the crawler’s copy.
example.com/services
Example Co<div id="app"></div>
543 tokens
example.com/services
Example Co — services, pricing, and coverage for regional operators. Our services span…
2,481 tokens
Where those two disagree, the difference is content no crawler can cite. A page below 80% is flagged; below 40% is critical.
Scores fall in three bands
A score is a reading, not a rank: it says how much of your site AI search can reach, read, and attribute today.
- 0–39Not AI-visibleAI search engines cannot reliably reach or read your main content.
- 40–69Partially visibleSome of your content is readable; specific pages or checks are holding the score down.
- 70–100AI-readyYour content is reachable, renderable, and structured for citation.
Five pillars, weighted
Each pillar is scored on its own checks and contributes its weight to the overall score.
- Crawler access20%
- Render visibility25%
- Structured data & entity20%
- Content extractability20%
- AI visibility15%
A free account runs the full technical scan
The lite scan reads two pillars on up to four pages. A free account scans all four technical pillars — crawler access, render visibility, structured data and entity, content extractability — and keeps the stored evidence behind every finding, so you can hand a developer the exact page, request, and response.
Every scan crawls politely, and you can opt out
Scans are bounded and say who they are. Standard requests identify as PeriscopeBot, robots.txt is honoured, and a domain that opts out is never fetched again.
- At most 60 requests per scan, spread across your site.
- Standard requests identify themselves honestly and never hide behind a browser user agent.
- robots.txt is fetched, parsed per user-agent group, and honoured.
- A per-domain cooldown stops repeat scans of the same site from piling on.
- Consumer AI apps are never scraped or automated — only your public pages are fetched.