A trust layer: verified, auditable, provably-forgettable memory around any AI.
chain aethelas-mainnet-1 · iPoS BFT consensus · ~5s blocks · WASM VM · 18/18 red-team attacks blocked
14 MB Rust binary · 273-weight continual learner · grounded knowledge engine · ZK proofs
52 modules · memory, reasoning, world models, planning, decentralized storage, autonomous loop
Aethelas is a trust layer: verified, auditable, provably-forgettable memory around any AI. It is not a foundation model, not general intelligence, and not a recursively self-improving system. Its benchmark publishes a 70% continual-learning score alongside its strengths, because a system whose value is trust reports its ceilings rather than hiding them.