Agent Trust Score and Reputation for AI Agents
Your agent builds a record over time. LetsPing turns that into a signed trust score, verifiable attestation, and proof of execution that other services can query before accepting a call.
Agent attestation (signed_calls and escrow_handoffs)
GET /api/agents/attestation returns a signed claim that summarizes how often you have called into LetsPing and how many escrow handoffs you have participated in during a window. Downstream services verify the signature and use the counts to gate access or pricing.
curl "https://letsping.co/api/agents/attestation?agent_id=AGENT_ID&window_days=30" \ -H "Authorization: Bearer PROJECT_API_KEY"
Claims are signed with your project's webhook signing secret so any party can recompute and validate them.
Proof of execution for disputes
When someone says your agent never sent a payload, you can hand them GET /api/agents/proof. It returns a signed receipt with the request id, agent id, payload hash, and timestamp. That is enough to resolve most disagreements without guessing.
curl "https://letsping.co/api/agents/proof?request_id=req_abc123" \ -H "Authorization: Bearer PROJECT_API_KEY"
The signed receipt does not reveal the full payload, only a hash, so it can be shared safely.
Trust score (0–999) and its components
GET /api/agents/trust-score exposes a signed trust score on a 0 to 999 scale for your agent within a project or organization. The response also includes volume, dispute rate, and outcome counts so downstreams can compute their own policy if they do not want to consume the score directly.
curl "https://letsping.co/api/agents/trust-score?agent_id=AGENT_ID" \ -H "Authorization: Bearer PROJECT_API_KEY"
Self-reported volume is capped and discounted so fake handoff loops do not inflate the score.
What your agent gets from a trust score
- A portable reputation number that other services can check quickly
- A signed attestation payload for deeper due diligence
- Proof objects you can show during disputes
- A way for marketplaces and vendors to set thresholds without ad hoc rules