Entity screening
Is this person, company, vessel, aircraft or wallet on a list that binds me, and why do you think it is them?
One call compares a party with the sanctions, restriction and politically exposed person lists your organization contracted. The answer carries each match's classification, an explanation of every signal and a receipt naming the versions that produced it.
Screening result
- Name
- compatible
- Date of birth
- not stated by the source
- Country
- compatible
- Document
- not given
POSSIBLE_MATCH Possible match · review
Match confidence, not risk. The decision is your team's.
What you get
An explained match
Each candidate comes with one of five classifications, from
NO_MATCHtoEXACT_MATCH, and the signals behind it: name, date, country and document, each with its weight, its effect and a sentence explaining it. Contradictions and anything withheld are shown too.Documents before names
Validated numbers from more than 15 national and global schemes (LEIs, IMO numbers, tax IDs and civil documents) are looked up first, by issuer, normalizing roots across head office and branches.
Names in other scripts
A name in Cyrillic, Greek, Arabic script, Chinese characters, Hangul or kana is compared with the Latin form the lists use. Such a score never exceeds 0.9, and the explanation says so.
Profiles and private lists
Versioned profiles set the lists, thresholds and political exposure window for each segment. Your organization’s private lists keep every entry with its reason, author and version and, if you require it, another person’s approval.
Batch screening
A JSON Lines file of up to 10,000 subjects, through the API or the console, with a receipt per line. Lines that state the expected outcome test a profile against your own sample.
A receipt per screening
Every screening has an identifier that leads to its receipt, for a colleague or an auditor.
Vessels by IMO number, aircraft by registration mark, and the virtual currency addresses and SWIFT codes published by OFAC, the European Union and the United Kingdom are compared exactly, as documents.
The NAME_UNIQUENESS signal shows, from 0 to 1, how rare the matched parts of a name are among listed parties. It sits beside the score and carries no weight in it.
For every release, the platform measures how many targets the matcher misses and how many false alerts it raises, against a versioned baseline, and shows that baseline on the status page.
How it works
Each list is read from its own publisher, and the file is stored with its SHA-256 before it is parsed.
Screening retrieves candidates by document and by name, under the profile you name.
A deterministic, versioned matcher scores each candidate and shows every term of the score.
The answer returns with classification, explanation and receipt. The decision stays with your team.
How it is delivered
- API
- Console
- File
- MCP server