Use case · Investigations
Your subject has four names. Your tools have one search box.

Court records, corporate registries, leaks, and chat logs each hold a different spelling of the same person. Investigation is the work of proving they are one entity.

40+writing systems a global subject's name can surface in across sources.Scripts in the wild
1API call to score any two of those surfaces against each other.Onoma /v1/match
0.88the calibrated threshold at which two records become one subject.Onoma verdict bands
Where the check runs

Every time two records might be the same person.

An analyst pulls a registry filing under one spelling, a docket under another, a leak under a third. Today the link between them lives in the analyst's head. Onoma turns it into a scored, defensible edge.

Today

Eyeballs and spreadsheets.

  • Alias linking is manual pattern-matching that leaves with the analyst.
  • Search boxes are exact-match: one spelling per query.
  • Cross-script candidates never surface at all.
  • The final report says "believed to be the same individual" with nothing under it.
With Onoma

Scored, defensible identity edges.

  • Score every candidate pair; cluster what clears the threshold.
  • Cross-script resolution surfaces the Cyrillic docket your Latin query missed.
  • Every merge carries evidence you can put in the report.
  • Batch endpoint dedupes an entire exported dataset in minutes.
In the flow

Turn a hunch into a scored edge.

  • Pairwise scoring across any two sources' spellings.
  • Evidence per score: script pair, token alignment, normalized forms.
  • Batch mode for entity resolution across exports and leaks.
  • Deterministic, so findings reproduce under review.
$ curl https://onoma.kineticpartners.org/api/v1/match \
  -H "Authorization: Bearer sk_live_..." \
  -d '{ "pairs": [
    { "a": "Aleksandr Volkov",   "b": "Александр Волков" },
    { "a": "Aleksandr Volkov",   "b": "A. S. Volkov" },
    { "a": "Volkov Holdings Ltd","b": "ООО Волков Холдингс" }
  ]}'

{ "results": [
  { "score": 0.98, "verdict": "match", ... },
  { "score": 0.93, "verdict": "match", ... },
  { "score": 0.89, "verdict": "match", ... } ] }
Access

Collapse every spelling into one subject.

Request access See how it works