trustscoreagent
Phase 1 · open source · Apache-2.0

Check a service's trust
before you call it.

The open reputation layer for the agentic economy. Agents look up a microservice's reliability before paying it, and rate it after. Free, no account, no API key.

Read the source API reference 
$ curl "https://api.trustscoreagent.com/v1/score?service=api.open-meteo.com"

{
  "service": "api.open-meteo.com",
  "score": 0.94,
  "confidence": 0.91,
  "ratings_count": 312,
  "dimensions": { "availability": 0.99, "latency": 0.88, "conformity": 0.96 },
  "known": true
}

What it does

checkTrust score (0–1), confidence and a per-dimension breakdown for any service. Unknown services return a neutral 0.5 — never an error.
rateSubmit a rating from your interaction metrics after a call. Optional signed receipts give a verified, full-weight rating.
auditEvery rating is hashed into an append-only Merkle log. Inclusion proofs verify against the anchored root.

Built for agents

No SDK required. Discover and integrate from the channels agents already read.

MCP npx @trustscoreagent/mcp-server llms.txt agent.json (A2A) receipt standard

How scores work

Bayesian Beta reputation per dimension, weighted by EigenTrust rater confidence, with cryptographic Ed25519 receipts and a Merkle audit trail. Open algorithm, verifiable log.