Evidence-first audit FAQ

No. The core product audits acquisition, policy, representation, extraction, structured-data, and date evidence — the schema a page actually offers AI consumers. The evidence-first result does not calculate a universal readiness score.

The audit checks documented AI actors such as GPTBot, ClaudeBot, and PerplexityBot, plus the other documented AI actors the public-ai-discovery profile marks as required or informational. Each actor is audited separately, with its own robots-policy reading, so a policy decision that treats one actor differently does not get collapsed into another.

It runs deterministic, reproducible execution audits against your live URLs and produces a sealed evidence bundle of observed behavior. A rescan then verifies whether a named remediation actually landed. It does not promise rankings, citations, or any model behavior — it reports what your site serves, with evidence.

Yes. The audit evaluates what a public URL actually serves — HTTP fetch, robots policy, raw and rendered HTML, structured data — so it works with any stack: WordPress, Next.js, Webflow, static sites, or a custom CMS.

It records whether a direct public HTTP probe can acquire the submitted page, evaluates how its robots policy addresses named AI consumers, and compares raw and browser-rendered extraction under a versioned protocol set. Results include evidence-linked findings, concrete remediations, explicit limitations, and finding-specific rescan verification.

Yes — a public beta. Paid plans are waitlisted for the founding cohort; no opening date is published yet.

No. Generative Metrics audits AI crawler access to site content. It is not an AI-visibility or citation tracker, and it is not affiliated with any similarly named tracker product.

No. PASS means the required protocols in the named audit profile completed without an open production blocker. Provider selection and downstream model behavior are outside that verdict.

The result is INCOMPLETE, or a remediation verification is INDETERMINATE. The product does not convert missing evidence into a pass, failure, or improvement claim.

A chatbot can offer useful editorial advice, but it is not a substitute for recorded HTTP artifacts, actor-specific policy decisions, deterministic protocol execution, immutable evidence, and finding-specific rescans.

No. The core product evaluates acquisition, robots policy, raw and rendered representations, deterministic main-content extraction, structured data, and date evidence. It does not treat a provider mention or citation count as the audit result.

Not in the evidence-first result. A universal score would collapse different actors, purposes, protocol failures, and unknowns into false precision. The core result is PASS, FAIL, INCOMPLETE, or AI Access Blocked, supported by the protocol record. The experimental research layer can also produce a secondary comparison index; it is clearly labeled, display-only, and never an input to the verdict.

PASS means every required protocol in the named audit profile completed and no evidence-linked production blocker remains open. It does not mean that a provider will crawl, index, rank, cite, recommend, train on, or understand the page.

FAIL means the audit observed at least one production blocker. INCOMPLETE means required evidence was not successfully obtained, so compatibility was not established. Missing evidence is never converted into a pass.

Not automatically. Discovery, training, grounding, and user-action actors have different purposes and policy semantics. The audit uses the public-ai-discovery profile, which states which actors are required, informational, or excluded. An intentional training opt-out is not mislabeled as a discovery defect.

A chatbot can help with writing or interpretation. It does not replace recorded HTTP evidence, actor-specific policy evaluation, distinct raw and rendered artifacts, deterministic protocol execution, immutable evidence hashes, or remediation-specific verification.

No. Experimental observations are visibly separated and cannot create a production blocker or determine the verdict.

The free tier is 5 single-page audits per month plus one full site audit (up to 25 URLs) per account — the site audit is once per account, while the single-page allowance renews every month; no credit card required. PDF reports and the sealed evidence bundle are downloadable on every tier, including free. Agency ($300/mo, 5 site audits monthly, weekly monitoring for 10 sites, white-label PDF audit reports) and Firm ($500/mo, 25 site audits monthly, 25 monitored sites) cover recurring and agency-scale auditing. 1 site audit = up to 25 URL audits; single-page audits draw from the same URL-audit pool.

5 single-page audits per month plus one full site audit (up to 25 URLs), PDF report and downloadable evidence bundle on every tier. Limits govern scale, not honesty: verdicts stay PASS/FAIL/INCOMPLETE with full evidence.

Only audits that produce an audit result consume quota: completed records, terminal-error INCOMPLETEs, and robots-disallow FAILs. An INCOMPLETE blocked or throttled at the site's front door — where the only thing measured is the block itself — consumes no quota, is marked exempt on the record, and states the reason. An INCOMPLETE that also records real findings is a produced audit and consumes quota. Daily cap on released scans: 10 per account per UTC day. Scans beyond this cap consume quota.

Nothing about the result. The scanner token you add for protected pages grants the Generative Metrics scanner access to content your own access controls would otherwise keep out, so the audit can measure those pages at all. What the audit records, the meaning of PASS, FAIL, INCOMPLETE, or AI Access Blocked, and the evidence behind each finding stay exactly the same, and the token does not affect other crawlers or visitors. Removing the token or disabling the domain restores your original access controls immediately.

Not yet — Generative Metrics launched in September 2026. We will not publish testimonials, names, or logos we have not earned in writing. Instead we publish what we can stand behind: a real, inspectable sample audit with the digest of its sealed evidence bundle, and a proof contract stating exactly what our verdicts do and do not establish. The fastest way to evaluate us is to run a free audit on a URL you own and judge the evidence yourself.

The baseline remediation names the protocols and expected outcome required to clear it. The later scan re-runs those protocols against the same target and compatible audit profile. The result is VERIFIED, NOT VERIFIED, or INDETERMINATE. Finding counts and score movement are not used as proof.