Legal

What a verdict is

A verdict from this API is a set of findings. Each finding names what was observed, which source it came from, and a link to that source. A risk score is derived from those findings under a named, versioned ruleset — it is never the primary artifact, and it is never bought from a vendor and resold as our own.

Every verdict is pinned to the ruleset version, the label snapshot and the chain height it was computed against, so the same verdict can be re-derived later and produce the same findings. You are meant to check our work, not to trust it.

What a verdict is not

A verdict is an input to your risk decision. It is not the decision.

You decide what to do with an address or a transaction. We tell you what we found and where it came from.

Your query data

When you check an address, the answer our label providers return about that address is recorded in our own label database and used in later checks — yours and other customers’.

What we record is the provider’s statement about the address: its category, the entity name the provider gives, the provider’s public page for it and the time of the answer. Your account, your identity and your check are never recorded with it.

We do not sell or share your query history, and we do not use your traffic to build a product that competes with your business. Your queries are retained only as the metering and audit records needed to bill you and to answer a dispute about a specific check.

Data sources and attribution

Our label layer is built from public sanctions data and openly licensed community datasets, and from the answers commercial label providers return for addresses checked through our service, each recorded with the provider and its public source. Some of them require attribution as a condition of redistribution, and this section is where that obligation is discharged.

Where a commercial data provider contributes to a verdict, we present findings derived from that provider’s data under the terms they granted us. We do not resell API access to any provider, and we do not redistribute any provider’s raw responses.

Provenance and corrections

Every label carries a dereferenceable source link, the provider that published it and the date it was observed. Labels are append-only: a label that is later contradicted is superseded and the earlier one remains readable, so a past verdict stays explainable. Nothing is deleted quietly.

If you believe a label about an address is wrong, tell us which address, which finding and why. We correct our own attribution by superseding it. We cannot correct a third party’s published designation — for those, the source link in the finding is where to go.

Availability and limits

Screening depends on chain data and on third-party sources, both of which can fail. When they do, you get a partial verdict that says so. Nothing in this document is a warranty that the service is available, complete or fit for a particular purpose, beyond what your contract with us says.