Why MetricDuck exists
A wrong number in finance isn't a rounding error — it's a bad decision with your name on it. MetricDuck is the SEC data layer I wanted when my own agents kept getting those numbers wrong.
I started by building financial-analysis agents and kept hitting the same wall: free APIs returned errors, missing metrics, and stale figures, while the reliable options were enterprise products — sales-gated, priced for institutions, and built for human terminals rather than agents.
The root cause is middlemen: most platforms aggregate from other aggregators, and errors creep in at every hop. So I built the opposite — a data layer sourced primarily from SEC EDGAR filings (the official record for US public companies), enriched with earnings-call transcripts, IR disclosures, and a market-data feed, with a link from every number back to the passage it came from. A human or an agent can verify any figure in one click.
It's fully automated, which is what lets it stay genuinely free and self-serve — no sales call, no card. Today that's 5,500+ US companies and 323+ pre-computed metrics, every figure traceable to its source. I build it in the open at @getmetricduck.
Questions, or spotted a data issue? I read every email.
[email protected]