Mnemom

The product is a trust and security infrastructure for AI agents that addresses liability and compliance when agents operate in regulated workflows by declaring intent, verifying reasoning before actions execute, anchoring governance on-chain, and enforcing cross-tenant defenses at gateways. It sells to developers, IT and security teams in B2B organizations deploying agentic systems. It is delivered as a five-layer runtime via infrastructure API and SaaS with signed Alignment Cards and Protection Cards as the core primitives.

Key features

• Signed Alignment Cards declare intent
• AIP verifies reasoning in flight
• CLPI governs lifecycle on-chain
• AEGIS signs cross-tenant defenses
• Four checkpoints per transaction
• Ed25519-signed Managed Rules
• Merkle-included hash-chained verdicts
• ZK-STARK proofs sampled at 10%
• Trust Ratings 0-1000 reputation
• Multi-agent coherence scoring
• Base L2 on-chain anchoring
• OpenTelemetry observability
• Safe House per-agent perimeter