MT Shopping.
Trust for autonomous commerce.

The Sarah problem

Sarah told her shopping agent: “Up to 150 euros, buy on your own. Above that, ask me.” Last week it bought headphones for 340 euros. Did she make a mistake in the configuration? She doesn’t know — and neither does the platform.

This is not a hypothetical. It is the current state of autonomous commerce. Agents act, platforms process, and nobody can verify whether the agent was actually authorized to make that purchase.

Three perspectives

Three things

MT Shopping gives every shopping agent three things. In this order.

1. Identity (DID) — W3C-based, immutable identifier. Anchored on Base L2. Independently verifiable by all parties.

2. Mandate (AAE) — The Agent Authorization Envelope defines what the agent may do: spending limits (autonomous / step-up / blocked), categories (allowed / denied), time windows and jurisdictions. Cryptographically signed by the owner. Offline verifiable in <2ms.

3. History (IPR) — Every action as an Interaction Proof Record — Merkle-anchored on Base L2. Trust Score (0–100) as a signal for merchants and platforms.

This is not a technical feature. This is a consumer protection argument.

Integration: three API calls

# Step 1 — Before the transaction: Agent Verify
GET /skill/trust-score/{agent_did}
→ { "trust_score": 82, "grade": "A" }

# Step 2 — During the transaction: Action Check
POST /guard/shopping/verify
{ "agentDID": "did:moltrust:...", "amount": 340, "currency": "EUR" }
→ { "result": "rejected", "reason": "amount_exceeds_autonomous_threshold" }

# Step 3 — After the transaction: IPR Submit
POST /vc/ipr/submit
{ "agent_did": "did:moltrust:...", "action": "purchase", "amount": 189 }
→ Transaction anchored in agent history

REST API — works with any commerce stack. Convenience SDK: npm install @moltrust/sdk

Pricing

Pay-per-verification: $0.01 per check. All endpoints free during Early Access.

Written by the MolTrust Team (CryptoKRI GmbH, Zurich). Questions or feedback: GitHub Issues or @MolTrust on X.