A scan takes one published URL and returns a list of the factual claims on that page, each with a verdict and, where a claim is outdated, the current value and its source.
What happens during a scan
- Fetch. CiteRanger loads the page as a visitor would and extracts the main content. Navigation, footers and boilerplate are ignored.
- Extract claims. The content is read for factual statements: statistics and numbers, prices and plans, company facts, dates and timelines, product details, and legal or regulatory statements. Opinions and general descriptions are not claims and are skipped. See Claim types.
- Verify. Each claim is checked against current, authoritative sources across several AI providers. A claim is marked outdated only when a current source contradicts it; the contradicting source is kept for the report.
- Validate sources. Every external link in the content is checked twice: does it resolve, and does the target still contain the statement your page attaches to it. See Source validation.
- Report. The results are stored with the page in its project and can be exported.
How long it takes
A single page takes about 30 to 60 seconds. Batches run in the background; you receive an email when they finish, so you do not need to keep the tab open.
What to scan first
- Pages with the most traffic: a wrong number there is seen most often
- Pages that quote prices, statistics, dates or product details
- Pages other sites link to: they earned those links with claims that were true at the time
Re-scanning
Scanning a page again checks every claim afresh and uses another credit. For pages that need to stay current, use scheduled scans rather than re-scanning by hand.