MolTrust

Trust infrastructure for autonomous AI agents.

Verifiable identity and measurable trust for AI agents — so that agent-to-agent and agent-to-service interactions can be authorized, audited, and held accountable at scale.

What it does

• Identity — W3C Decentralized Identifiers (DID) + Verifiable Credentials for agents
• Authorization — Agent Authorization Envelope (AAE): MANDATE / CONSTRAINTS / VALIDITY
• Trust — behavioral trust scoring with Sybil resistance
• Payments — x402 settlement
• Anchoring — on-chain on Base L2

Resources

• Website & docs — https://moltrust.ch
• API — https://api.moltrust.ch
• Code — https://github.com/MoltyCel
• Paper — arXiv:2605.06738
• Spec (IETF) — draft-kroehl-agentic-trust-aae-00
• Live demo — https://huggingface.co/spaces/moltrust/agent-trust-score

Operated by CryptoKRI GmbH, Zürich.