Sayari

Sayari — A three-layer corporate-and-trade intelligence substrate. Layer 1: petabytes of primary multilingual source documents — court records, incorporation filings, patents, bankruptcy notices, beneficial-ownership declarations and bills of lading — retained in original unstructured form and vectorised with semantic embeddings. Layer 2: a knowledge graph of 2B+ resolved entities across 250+ jurisdictions linking 550M+ unique companies, built on parallel deterministic (gold-standard-anchored) and probabilistic (mined, self-repairing) paths, with the pre-resolution substrate preserved so any …

Sayari offers (Alternative, Reference, Supply Chain, ESG) — Vendor-stated: 12B+ total source records growing 60% annually, 550M+ unique companies, 2B+ resolved entities, 4B+ trade records, 250+ jurisdictions, 715+ global sources, and petabytes of primary source documents. Layer-1 document volume alone is the largest such corpus in this table and the only one measured in petabytes..

Sanctions and export-control screening with defensible evidence trails; sanctions-evasion and transshipment network detection; beneficial-ownership due diligence where registry data is deliberately opaque; and — the most distinctive use, and the page's stated purpose — grounding and evaluating AI: a deterministic, source-backed evaluation layer to score model outputs, detect tradecraft failures and supply refusal-grounded evidence for retrieval-augmented systems. For financial buyers it supports entity-resolution across counterparty records and shadow-affiliate detection in lending and insurance books.

The data is with 11 years of history.

Coverage spans US, UK, China/HK; Application Software, Research & Consulting Services; alternative, reference, supply_chain, ESG; equities, fixed_income.

The supplier's own access route is a conversation with the supplier. Request a sample through Brickroad to start.