Finds the hole, fixes it and proves it by measuring before and after
A visibility machine: it checks the technical state, whether AI engines cite you and what content is missing, records each finding with its source and time, writes the fix and — after the release — measures the same page again. Every finding carries a hash, every proof has a before and an after; a claim without a measurement shows up in the cockpit as “waiting for proof”.
The problem it solves
- A report full of recommendations, and nobody checks afterwards whether the fix was actually released.
- AI search engines do not cite the site — and nobody measures how often, or on which questions.
- “Fixed” is claimed, not proven: there is no before and no after number, so the work cannot be billed on results.
What you get
A finding with evidence
Source (PageSpeed, crawl, GA4, GEO check), time and a server-computed hash. Without it a finding does not enter the report.
The fix in the cockpit
A draft fix for every finding; the release goes through your flow (repo or SSH) with your approval — the machine does not publish on its own.
Before and after proof
The same page is measured again. The result is a pair of numbers with a hash, e.g. LCP 4.8 s → 2.1 s — not “improved”.
The client report
Findings, proofs and the change in numbers for the period; the existing report templates and monthly report are part of the machine.
What it looks like
