Don’t trust it. Check it.
Engram is your AI’s long-term memory, owned by you — portable, private, and provable. Every operation leaves a value-blind receipt anyone can verify offline, down to proof a deleted fact can no longer be recalled.
Your memory.
Not their black box.
An AI’s long-term memory normally lives in someone else’s black box, so you trust it by promise. Engram makes it yours — portable across models, private to your keys, and provable.
Every operation
leaves a proof.
A value-blind receipt anyone can verify offline — no server, no trust required.
- Custody
- Grounding
- Completeness
- Consent
- Data-minimization
- Forgetting
Prove it
forgot.
Delete a fact and Engram proves it is unreachable in the deductive closure of the signed survivors, under a fixed rule-set R — a deletion you can check, not just trust.
Scoped honestly: unreachable under rule-set R — not a claim of cryptographic erasure. 26 adversarial attacks held.
One zero-dependency library, independently red-teamed. Open source. No token.
Own what
it remembers.
Value-blind facts under classical discrete-log. Not FHE, not post-quantum, not audited for deployment. No token, ever. © 2026 Engram · open source.