Scan limitations

VisibleMax measures AI visibility from public signals. Understanding the limits below helps you interpret scores, empty findings, and “not mentioned” results without treating them as bugs.

Public HTML focus

• We crawl publicly reachable pages over HTTPS/HTTP.
• We do not log into customer accounts, bypass paywalls, or scrape private APIs.
• Content that exists only inside authenticated apps is invisible to the scan.

No private model weights

VisibleMax does not inspect or reverse-engineer private model weights. Probes use provider APIs and public assistant behavior. Training-data membership for any given page cannot be proven from the outside.

Crawl depth and page caps

• Discovery follows links and common entry points from the URL you connect, up to a maximum pages-per-scan cap (currently on the order of tens of pages, not a full site mirror).
• Deep archives, orphan pages, and unlinked PDFs may never be fetched.
• Time budgets apply — very slow hosts can finish with fewer pages crawled.
• Duplicate URLs, redirects, and blocked paths reduce effective coverage.

Probe and analysis limits

• Only key-gated providers with working credentials are probed.
• Prompt sets are representative buyer questions, not an exhaustive query universe.
• Structural analysis reasons only over the crawl corpus — it will not invent pages you did not crawl.
• Plan entitlements may gate how much of a report (recommendations, history, competitors) you can see.

What we never claim

• That we control ChatGPT, Claude, Gemini, or any other vendor’s ranking or training pipeline.
• That fixing one finding will force a mention overnight.
• That a scan equals a full SEO crawl of every URL on your domain.
• That private or authenticated content was evaluated.

FAQ