Scanning & results
Input modes
- Paste text — one text, instant feedback with inline highlighting. Ctrl/⌘+Enter scans.
- Upload files — batch of .txt / .md files, one verdict per file.
- Upload Excel / CSV — full catalogues: you map the content column (plus optional ID column), every row is scanned, and the export returns your original file with verdict columns appended.
Scan depths
| Depth | What runs | When to use |
|---|---|---|
| Fast | Keyword & regex layer only — instant | Quick sweeps for obvious violations |
| Standard | Keywords + AI classification | Regular content review (default) |
| Advanced | Standard + legal evidence retrieval when issues are found | Balanced depth and cost |
| Deep | All layers including evidence for every finding | Pre-launch review of new copy |
Severity levels
- Violation — the scan matched an unauthorized health claim (e.g. disease claims under Art. 14 HCVO, court-confirmed terms like "Detox"). Strong signal to remove or reformulate.
- Warning — borderline: implicit health promises, traditional-use framing, general well-being claims (Art. 10(3) HCVO). Deserves human review.
- Pass — an authorized claim used correctly per the (EU) 432/2012 positive list.
- Clean — no health claim detected.
Suggested rewrites
With Rewrite flagged claims enabled, VeriClaim produces a minimally-edited draft for each flagged item: violating phrases are removed or replaced sentence-by-sentence, and the result is re-checked against the keyword layer. The rewrite is a suggestion — review it before use.
Exports
Every scan exports to Excel with one row per finding (flagged text, severity, regulation reference, suggested replacement). Spreadsheet scans export your original file augmented with verdict columns.