VeritasEngine, explained for beginners

VeritasEngine checks an AI agent's output with fixed rules; one failed check and the output is refused.

What it is

AI agents produce output that other systems then trust: a JSON record, a number, a quote from a document, a file. VeritasEngine checks that output before anything trusts it, using rules that give the same answer every time, with no AI model involved in the checking.

You list the checks: is it valid against this schema, is this arithmetic exact, does this file have this fingerprint, does this quote really appear in the source, is this forbidden text absent. The answer is VERIFIED only if every check passes. One failure makes the whole output REFUTED; a check it cannot decide makes it UNDECIDABLE, never VERIFIED. Every result comes with a signed receipt.

An everyday comparison

A customs inspection with a checklist: one item wrong and the shipment doesn't get in.

How it works, in 4 steps

1. Send the agent's output and the checks you want.
2. Each check runs with fixed rules, no AI.
3. All pass: VERIFIED. Any fail: REFUTED. Can't decide: UNDECIDABLE.
4. A signed receipt records the result.

Try it now

Open the five live demo tools → Every button sends a real request to the live product and shows the real answer.

See it on the homepage: VeritasEngine section →

Who it is for

Teams whose agents produce records, numbers, quotes or files that other systems act on.

FIRME AI CORPORATION · CAIN-42 is a Byzantine trust fabric. Numbers marked "our own check" were run by us, not by an independent party. Nothing on these pages is a certification.