Sourced, resolved, cited.
WhiteIntel fuses 24 public and licensed sources into one graph of 37.1M+ entities — every claim carries the record behind it. Here's exactly where the data comes from, how the same entity is resolved across sources, and how it's secured.
Five ledgers, one identityThe same entity, across sources
- 1 · Identifier join — exact strong-id match (LEI, sanctions id, company number, NIP, UEN, SEC CIK).
- 2 · Name-block — name + jurisdiction, only where a block spans ≥2 registries.
- 3 · Sanctions bridge — links null-jurisdiction sanctioned parties onto same-name registry/leak records.
Matches collapse into one cluster_id; the dossier shows every record in the cluster.
Locked down by construction
- Row-Level Security on every table; the public API reads through a locked-down service role.
- Input validation + rate limiting on every public endpoint; write functions execute-revoked from anon.
- No secrets shipped to the client. Enforced security headers. Posture independently audited.
Every entity, edge and risk flag links to the registry or leak record behind it. Nothing is asserted without provenance.
A missing ownership edge means “not yet observed in our sources”, not “does not exist”. We never imply completeness we don't have.
Risk levels are heuristic investigative leads. WhiteIntel is not a legal determination of beneficial ownership.
Where an entity isn't yet risk-scored we say “Unscored” rather than defaulting it to low.
Questions, answered.
See it for yourself.
Open any entity — every claim on the dossier links back to its source.