Skip to content

Grade us. Then run it again yourself.

Two public benchmarks run live against the very engine that serves every match. Nothing is special-cased for the test. Results cache per reference pack and every run returns a sha256 receipt so two parties can compare numbers without trusting either one.

The public sets, live.

A representative CRM file shows field performance: realistic noise with web domains at real-world fill. An adversarial set shows the honest floor: built to break us with no domains to match on. Both are public rows built from CC0 GLEIF data.

The numbers report as they are. A close call routed to review counts as neither right nor wrong. An honest miss stays a miss. The wrong-entity rate prints first because a confident wrong answer is the one that costs you.

Point an agent at the same endpoint or press the button. Run it again on the same pack and the receipt lands identical.

GET /api/eval/publicLive

Run both public benchmarks yourself, live against the serving pack. No key needed. A representative CRM file shows field performance and an adversarial set shows the honest floor. Results cache per reference pack so every run returns the identical receipts.

The keyholdout lift.

We hold out records whose identifying keys the matcher never sees and ask it to find them anyway. Measured. Not projected.

−37%
Unmatched records · keyholdout

Matched 92.0% on v8.10 and 95.0% on the v8.11 deployed arm: unmatched fell 8.0% to 5.0%.

90%92%94%96%98%100%92.0% · v8.1095.0% · v8.11 deployed
Matched rate · held-out key set · axis 90–100

What a source of record is.

By sources of record we mean distinct underlying registration authorities witnessed on our ingested records: the company registers, filing systems, and authority codes our records actually carry. Counted and never estimated.

Measured: 937 distinct registration authorities counted against the 2026-07-14 GLEIF golden copy with format-verified codes and sentinels excluded. Direct feeds and referenced registries only grow the union. Where the site rounds it rounds down.

The landing ledger binds the served count from /api/stats and activates only when the api serves it. The coverage atlas catalogues the pull-sources jurisdiction by jurisdiction.