How it works
From scan run to signal history.
CertScore.ai is built around a narrow output model: scan metadata, structured signals, summaries, and changes over time.
1Crawl a bounded page set
Each scan starts from the homepage, discovers links, and selects a limited number of pages based on the current plan.
2Run technical checks
The scanner runs automated accessibility checks, detects tracker activity, looks for common policy pages, and notes disclosure-related cues.
3Store derived signals
The system stores signal values, counts, timestamps, and change events. It does not store raw HTML, page content, screenshots or personal data.
4Compare against the prior scan
When a previous completed scan exists, CertScore.ai records added, removed, and changed signals so recent changes are easy to review.
