How it works · AI Attestation
How AI Attestation works.
A signed record of what a model was asked, what it answered, and which model it was. This page covers the mechanism: what happens to a request, what is signed, and how someone outside your company checks it.
AI attestations, live
Live
The path of a request
- Call your model as usualInvoance sits beside the call, not in front of it
- Post the exchangeInput, output, provider, model name and version
- Hash and signinput_hash, output_hash and payload_hash are computed, then the record is signed
- Verify laterAnyone holding the text can confirm it matches the attested hashes
Step by step
Call your model as usual
Invoance sits beside the call, not in front of it. No proxy, no added latency on the model path.
Post the exchange
Input, output, provider, model name and version.
Hash and sign
input_hash, output_hash and payload_hash are computed, then the record is signed.
Verify later
Anyone holding the text can confirm it matches the attested hashes.
Technical factsAPI reference
| Hashes | input_hash, output_hash, payload_hash (SHA-256) |
|---|---|
| Signature | Ed25519, per-organization key |
| Model context | provider, name and version are part of the signed record |
| Verify | POST /v1/ai/attestations/{id}/verify |
| Providers | Any. The API takes text, not credentials |
How a third party verifies
Verification does not need an Invoance account or access to your systems. The verifier recomputes the hash from what they hold and compares it with the signed record.
- Opens the proof linkNo account, no dashboard access
- Recompute the hashSubmitted payload against anchored hash
- match_result: truePlus issuer domain and signing time
Try it against a real record.
Open the verifier, or create a record of your own on the free plan.