About BalanceProof

BalanceProof is the verification layer over SEC EDGAR balance sheets. Each filing is checked against Assets = Liabilities + Equity when it reports the lines needed; where it does not, it is marked not testable instead of passing. Every figure is pulled as filed, and the check runs before it is stored. When a filing does not balance, the page says so with the reason instead of adjusting the numbers. 6,257 companies covered. 5,645 reconcile directly. 198 flagged with a reason. 0 hidden.

Who builds it

BalanceProof is built in Ottawa, Canada, by founder Dominique Church. It started as a tool for a personal project and became a public API because every free SEC data source we tried disagreed with the filing somewhere.

Who it is for

We build it for data engineers, quants, fintech teams and analysts who work with US public companies, in the US and Canada. Coverage is the companies that file periodic reports (10-K and 10-Q) with the US SEC. Our focus is one thing: checking SEC balance sheets against Assets = Liabilities + Equity and publishing the reason when a filing does not balance. We write about it in our research: bank balance sheets, XBRL duplicate tags, restatements and point-in-time figures.

The method

Each filing is checked against Assets = Liabilities + Equity when it reports the lines needed; where it does not, it is marked not testable instead of passing. A filing that reconciles is published with its figures. A filing that does not is published with the reason: noncontrolling interests, mezzanine equity, rounding, or a genuinely broken filing. It is never quietly adjusted until it balances. The counts are public and the categories are named: how we check the numbers. How many filings fail and why is in the report on 38,568 balance sheets.

What it does not do

No predictions. No scores. No recommendations. Nothing is estimated, smoothed or restated. Where a filer does not report a component, the page says so rather than showing a zero.

Using it

Looking companies up on this site is free and needs no account. The machine-readable version, the REST API and the bulk CSV, is what costs money: what it costs, and the API reference.