Open-source trust and governance layer that verifies, signs and monitors AI agents in regulated workflows.
What it does
Mnemom is a trust and governance runtime for AI agents. It layers an Agent Alignment Protocol (signed intent declarations), an Agent Integrity Protocol (verifying reasoning mid-run), on-chain lifecycle governance and a cross-tenant defensive network. Every verdict is cryptographically signed and hash-chained, giving teams audit-ready oversight when deploying agents in regulated settings. The core protocols are open source under Apache 2.0.
Core features
Agent Alignment Protocol publishing signed intent cards
Agent Integrity Protocol checking reasoning before actions execute
On-chain card lifecycle governance with observability
Cross-tenant defensive network (AEGIS/Safe House)
Ed25519-signed, hash-chained, auditable verdicts
Portable cryptographic trust ratings for agents
Best for
→Governing AI agents in regulated workflows
→Producing audit-ready compliance evidence for agents
→Screening agent transactions for policy violations
→Sharing threat rules across tenants in real time
Pricing
Developer
$30/mo
Team
$79/mo