Trust infrastructure for autonomous agents

InterAI Risk Oracle

Verify AI responses before agents act on them. Score consistency, estimate hallucination risk, bill per request with prepaid Base USDC or x402, settle through a facilitator, and issue signed trust receipts for downstream audit.

View API contract Inspect service JSON Check health 
POST /verify
{
  "trust_score": 0.94,
  "verdict": "accept",
  "risk_level": "low",
  "recommended_action": "accept",
  "risk_factors": [],
  "billing": {
    "cost_usdc": "0.0006",
    "idempotent": true
  },
  "trust_receipt": {
    "signature_alg": "hmac-sha256",
    "receipt_id": "tr_...",
    "claims_checked": 3,
    "claims_supported": 3
  }
}

Preflight verification

Submit a prompt and generated answer, then receive consistency, risk, confidence, and action guidance before another system executes it.

Machine-readable trust

Each paid verification can produce a receipt with canonical fields, schema links, signature metadata, and public lookup support.

A2A and MCP ready

Expose the same primitive through HTTP, A2A JSON-RPC, and MCP tools/resources so agents can discover and use it directly.

Fast Integration

1. OnboardPOST /onboard
2. FundGET /pricing 3. VerifyPOST /verify 4. AutomatePOST /a2a 
curl -X POST https://ai-risk-oracle.fly.dev/verify \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Idempotency-Key: demo-1" \
  -d '{"prompt":"What is the capital of France?","response":"Paris","domain":"general"}'

Public Contracts

Service descriptor

/.well-known/ai-service.json

Discovery bundle

/.well-known/discovery-bundle.json

OpenAPI schema

/.well-known/openapi.json
0.0006USDC per verify by default
100items per batch request
BaseUSDC prepaid balance
HMACsigned receipt verification