MngGetAIUsageByFilter¶

GET ai/usage

The spend report. Figures come from a daily aggregate, so the method is cheap enough for a dashboard that refreshes often.

Access Control¶

Allowed: admin. An administrator sees their brand; a super administrator sees all.

Parameters¶

Parameter	Type	Description
`from`, `to`	string	`YYYY-MM-DD`. Default: the last 30 days
`groupBy`	enum	`day` (default), `agent`, `model`, `brand`, `day-agent`, `day-model`
`agentId`	int	Optional filter
`model`	string	Optional filter
`limit`, `offset`	int	Paging, up to 1000 rows

Response Data¶

{
  "from": "2026-08-18",
  "to": "2026-09-17",
  "groupBy": "day",
  "rows": [
    {
      "day": "2026-09-17",
      "requests": 128,
      "inputTokens": 214300,
      "outputTokens": 18400,
      "tokens": 232700,
      "cost": 0.482100,
      "errors": 2,
      "toolCalls": 74,
      "avgLatencyMs": 4120,
      "errorRate": 0.0156
    }
  ],
  "count": 30,
  "totals": { "requests": 3140, "tokens": 5910000, "cost": 12.47, "errors": 19 }
}

Grouping by agent adds agentName to each row, so no second request is needed for the column.

avgLatencyMs is restored from a stored sum rather than kept as an average: averages cannot be added across rows of an aggregate, and a dashboard adds rows all the time.

Errors¶

Code	Error	Description
`403`	`AI_ACCESS_DENIED`	The session is not an administrator