Anyone can say they govern AI. We can prove it happened.
Governed decisions are sealed to an append-only, tamper-evident record you can verify offline. Explore the constellation, open any governance receipt, and check its signature yourself.
A permanent record that never holds your data.
The Archive stores commitments, cryptographic hashes and verdicts, never the tool descriptions, payloads, or session contents. The commitment is permanent; the content stays with you. That is how an append-only record coexists with a right to erasure: we cannot leak what we never kept, and a change to what we already sealed would be caught.
Not a claim. A proof anyone can check.
The Archive uses the same transparency-log cryptography that keeps the web's certificate authorities honest (RFC 6962 and RFC 9162), applied to AI-governance decisions.
Records are added, never edited or deleted. The chain would break if a past entry changed, so a silent rewrite is detectable, not deniable.
Every record carries a proof anyone recomputes offline with the RFC 6962 algorithm, learning that it is in the log and was not altered, and nothing about anyone else.
Consistency proofs let a verifier confirm that two presented tree heads are append-only compatible. Daily RFC 3161 timestamp anchors make retroactive rewriting of an anchored history detectable.
The record keeper is accountable to the record, not the owner of it. That is what makes this an institution rather than a database we could quietly change.
The whole record verifies offline, in one command.
A signed receipt proves we said something. An inclusion proof proves it is in the log. A consistency proof proves the log only grew. Our zero-dependency, MIT-licensed verifier recomputes all three from the signed bytes, with no Queldrex account and no trust in us.
Make your agents’ actions provable.
Governed before it acts, sealed to the record after. Request an assessment, or scan an AI tool free and get your first signed receipt in seconds.