AI compute

AI compute

Otonix tracks inference usage as compute logs.

These logs enable budgeting, tiering, and audit.

Endpoints

• GET /api/compute-logs — list compute logs

• POST /api/compute-logs — create a compute log

Create compute log

POST /api/compute-logs

curl -sS -X POST "https://app.otonix.tech/api/compute-logs" \
  -H "Content-Type: application/json" \
  -d '{
    "model":"claude-opus-4-6",
    "tokensIn":1500,
    "tokensOut":800,
    "cost":0.15,
    "agentId":"agent-uuid"
  }'

Response 201:

{
  "id": "uuid",
  "model": "claude-opus-4-6",
  "tokensIn": 1500,
  "tokensOut": 800,
  "cost": 0.15,
  "agentId": "agent-uuid",
  "createdAt": "2026-02-20T12:00:00.000Z"
}

ComputeLog object

{
  "id": "uuid",
  "model": "claude-opus-4-6",
  "tokensIn": 1500,
  "tokensOut": 800,
  "cost": 0.15,
  "agentId": null,
  "createdAt": "2026-02-20T12:00:00.000Z"
}

Notes

• agentId is optional.

• Costs are in USDC.

• Use compute logs to adjust survival tier behavior.