| Name screened in one jurisdiction Charged once for each proposed name that is actually screened against one jurisdiction's registry — visual and phonetic similarity against that office's curated famous-mark seed dataset, scoped to your Nice classes, returning a clear / caution / high-risk collision verdict with the matching marks. One charge per name per screened jurisdiction, so three jurisdictions for one name is three charges. Nothing is charged when a jurisdiction returns not_screened (no data source enabled, an unscreenable name, or a registry source error), and duplicate names are screened and charged only once. Informational screening only — not legal advice. | |
| Name generated and screened Fires once for each name generated from your positioning brief that survives the deterministic filter (profanity, your banned words, duplicates, and your must-include / avoid rules) and is then screened for trademark collision. Generate mode carries no charge — you supply your own LLM key, so there is no inference cost to pass on. The event still fires so your run's charge breakdown shows the work performed. One event per candidate covering every jurisdiction you selected. Candidates rejected by the filter are listed in the output at no charge, and a candidate that could not be screened anywhere emits nothing. Informational screening only — not legal advice. | |