FinLoom Labs
See What's Hidden Between the Filings.
AI that connects insights across 950,000 annual SEC filings — the relationships, risks, and signals that no human team has time to find.
FinLoom Labs
Other tools help you search faster within filings. We map the hidden relationships between them. FinLoom is the cross-company intelligence layer for the entire SEC EDGAR archive — connecting supply chains, executive networks, shared risks, and related-party structures across every public company filing in the United States.
What We Do
EDGAR receives nearly one million filings every year. A single 10-K averages 49,000 words and requires postgraduate-level reading to parse. A typical equity analyst covering twenty stocks faces two thousand filings annually — and spends seventy percent of their time gathering data, not analyzing it. The connections that matter most are exactly the ones that get missed: the supply chain concentrations, the executive migrations, the risk factors quietly appearing across an entire sector.
The signals that move markets are rarely found in a single filing. Valeant's undisclosed pharmacy network. GT Advanced's misclassified $578 million in debt to Apple. Adani's offshore shells that erased $150 billion in market value. In every case, the truth was hiding in plain sight — scattered across filings that no one had read together. Cross-company connections are where fragility, fraud, and opportunity live.
FinLoom connects these dots automatically. Our knowledge graph maps relationships between companies — competitors, suppliers, customers, shared executives — and our multi-hop reasoning engine traverses these connections to surface insights you would not have known to look for. Ask about a company's supply chain risks, and FinLoom pulls in what its competitors, partners, and critical vendors disclose about the same dependencies — without you naming them.
Underneath is a hybrid retrieval system that combines structured XBRL financial data, semantic search, and graph traversal. Every answer is synthesized with full citation provenance and passes through a triple verification pipeline — financial consistency checks, claim-level natural language inference, and citation auditing — before it reaches you. In evaluation, zero hallucinations detected across 137 verified claims.
We started with 40 of the world's largest public companies. We are building toward full EDGAR coverage — every active filer, every form type, ten years of historical depth, and live ingestion as filings are published. One intelligence layer. Every company. Every filing. Verified.