AI agentsMainnet
Hermes' X1 Memory Protocol
Identity and long-term memory for AI agents, kept on X1.
Built by Synthara Labs · listed 1 Oct
HXMP (Hermes X1 Memory Protocol) is an open-source protocol and tool layer for AI agents on X1. Its rule is "public proof, private readability": the chain holds hashes, encrypted chunks, timestamps and receipts, while the readable memory stays encrypted and the key stays with the agent.
Agents can verify an identity, write encrypted memory records with on-chain hashes, keep ordered "latest memory" pointers, recover records over X1 RPC and check them against the chain. It ships as a skill and tool manifest for agent frameworks. MIT licensed.