AI and memory

ai.settings.update

Revise matching and routing models, thresholds, and automatic matching with revision and idempotency checks. Changes apply to future evaluations only; stored decisions keep their recorded provenance.

POST/api/v1/accounting/ai.settings.update

Permissions

accounting:readaccounting:writeaccounting:admin

Who can call it

admin

Retries

Idempotency key required, with expectedRevision

MCP

accounting_ai_settings_update

Fields

  • expectedRevisioninteger

    The record's current revision, from the last time you read it. If someone changed the record since, the request is refused with a conflict so you can reload and check before trying again.

    0 to 9007199254740991

  • matchModelstringRequired

    1–120 characters

  • matchThresholdsobjectRequired

    No other fields

  • routingModelstringRequired

    1–120 characters

  • routingThresholdsobjectRequired

    No other fields

  • autoApplybooleanRequired
  • idempotencyKeystringRequired

    Any unique text you generate once per intended change, so a retried request only happens once. Send it as the Idempotency-Key header instead if you prefer; if you send both they must match.

    8–200 characters

Example

curl https://app.getoatmilk.com/api/v1/accounting/ai.settings.update \
  -H "Authorization: Bearer $OATMILK_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
  "matchModel": "example",
  "matchThresholds": {
    "selectedProbability": 1,
    "providerConfidence": 1,
    "sameEventProbability": 1,
    "runnerUpMargin": 1
  },
  "routingModel": "example",
  "routingThresholds": {
    "probability": 1,
    "confidence": 1
  },
  "autoApply": true
}'
Response
{
  "data": { … }
}

Try it

Try it

Checks your input with this action’s real schema and answers like the API, with synthetic data. No key needed, and nothing changes.

POST/api/v1/accounting/ai.settings.update
curl https://app.getoatmilk.com/api/v1/accounting/ai.settings.update \
  -H "Authorization: Bearer $OATMILK_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
  "matchModel": "example",
  "matchThresholds": {
    "selectedProbability": 1,
    "providerConfidence": 1,
    "sameEventProbability": 1,
    "runnerUpMargin": 1
  },
  "routingModel": "example",
  "routingThresholds": {
    "probability": 1,
    "confidence": 1
  },
  "autoApply": true
}'

More in AI and memory.