Financial data built on primary sources
Synapse Discovery provides US equity, fund, and macroeconomic data to quantitative researchers, developers, and AI systems — assembled directly from regulatory filings and official statistical releases.
Computed from the production store, not marketing estimates.
What we provide
Equities
Daily OHLCV for the full US universe including delisted, bankrupt and acquired securities, with split and dividend adjustments. As-reported fundamentals parsed from SEC XBRL, retaining every filing vintage.
Funds
ETFs and mutual funds with full position-level holdings from SEC N-PORT — the complete portfolio as filed, not a top-ten snapshot — plus NAV history, SEC-sourced fees, and reported monthly returns and flows.
Macroeconomic series
Interest rates, inflation, output and employment sourced directly from the US Treasury, Federal Reserve statistical releases, the Bureau of Labor Statistics, the Bureau of Economic Analysis and the EIA, with point-in-time vintages for the major series.
Insider activity
SEC Forms 3, 4 and 5 resolved person-centrically across issuers, with Rule 10b5-1 plan status where the filing records it and amendments reconciled rather than guessed.
How we build it
Four commitments determine what the API returns. They are design constraints enforced in the pipeline, not aspirations.
Point-in-time by construction
Every fundamental carries the filing date on which it became public, and
every restatement is retained as its own vintage. An as_of
query returns what was actually knowable on that date — the cutoff is
applied to publication, never to the period being reported.
Survivorship-free
Securities that stopped trading remain queryable at their final state, and their history ends where trading ended. A universe reconstructed for a past date contains the companies that existed then, not the ones that happen to have survived.
Primary sources
Fundamentals come from SEC filings, fund holdings from N-PORT, insider activity from Forms 3/4/5, and macro series from the agencies that publish them. Where a figure is derived rather than filed, the response says so.
Stated coverage
Fields we cannot source are null, never imputed, and
meta.coverage
reports which are populated. Our classifications — sector, category, style
box, star rating and the SD2000 index — are computed in-house from public
filings and published as first-party classifications; they are not industry-standard schemes and are
never presented as such.
Access
A REST API, bulk Parquet exports, and a native
Model Context Protocol server
for AI agents, which can connect and begin querying without a key. Documentation
is at /docs; machine-readable
descriptions are served at /llms.txt
and /api/v1/openapi.json.
Known gaps and limitations are published on Known limitations; terms of use are set out in our Terms of Service.
Evaluate it against your own data
The free tier is sufficient to reproduce any coverage or methodology claim on this page.