Tyche Institute
Open standards and reference implementations for verifiable AI agent attestation.
Tyche Institute is an Estonian non-profit research entity (in formation) advancing open standards, reference implementations, and reflective inquiry at the intersection of cryptography, AI governance, and the philosophical questions that surround them. Named after Τύχη, the Greek goddess of fortune — the classical embodiment of randomness that grounds modern cryptography.
What we do
Tyche Institute maintains and develops EATF (Agent Trust Framework) — cryptographic infrastructure for verifiable AI agent attestation. We publish open research, contribute to standards bodies, and provide reference implementations.
-
Open by design.
All code under Apache 2.0. All specifications open for community review.
-
Verifiable by structure.
Cryptographic evidence over claimed compliance. Reference implementations that auditors can verify.
-
Independent by intent.
Non-profit operation. No commercial dependencies.
Flagship
EATF — Agent Trust Framework
EATF is open-source infrastructure for cryptographically verifiable AI agent attestation. It produces signed, timestamped, offline-verifiable evidence packages that address record-keeping, transparency, human oversight, and post-market monitoring requirements under emerging EU regulation.
Active deployments demonstrate EATF in regulated sectors: education, environmental data, construction.