Usage and analytics - ThinnestAI

Four read-only endpoints for building your own dashboard. Every number is the same one the console shows — the Usage page, the dashboard, and each agent’s Analytics page — so your dashboard and the console always agree. A Read-only key is enough for all of them. See Choose what each key can do.

Method	Path	What
`GET`	`/usage`	Plan, replies this month, allowance, balance, replies per day
`GET`	`/agents/{id}/analytics`	One agent’s conversations, leads, escalations, unanswered questions
`GET`	`/calls/summary`	Call totals and a series
`GET`	`/whatsapp/summary`	WhatsApp sent, delivered, read, failed, received

Money is in micro-units of currency: 1000000 is one rupee (INR) or one dollar (USD). Divide by 1,000,000 to show it.

Usage and balance

GET /api/v1/usage?days=30

{
  "plan": "payg",
  "currency": "INR",
  "period": { "start": "2026-10-01T00:00:00.000Z", "end": "2026-11-01T00:00:00.000Z" },
  "replies": { "used": 1240, "included": null, "metered": true, "stopped": false },
  "wallet": {
    "balanceMicro": 1850000000,
    "availableMicro": 1350000000,
    "heldMicro": 500000000,
    "creditMicro": 0,
    "creditExpiresAt": null
  },
  "daily": [{ "day": "2026-10-04", "replies": 52 }, { "day": "2026-10-05", "replies": 61 }]
}

Field	Means
`period`	This calendar month, in UTC — what `replies.used` counts
`replies.metered`	`true`: every reply is paid from your balance. `false`: an allowance of `included` replies this period, and `stopped` once it is used up
`wallet.balanceMicro`	Your balance, as on the billing page
`wallet.availableMicro`	What can be spent now — the balance less money held for phone-number rent renewing soon
`wallet.creditMicro`	Plan credit included in the balance, and when it expires
`wallet`	`null` on the Free plan, which has no balance
`daily`	Replies per UTC day for the last `days` (1–90, default 30)

Agent analytics

GET /api/v1/agents/{id}/analytics?days=30&rows=20

{
  "days": 30,
  "totals": {
    "conversations": 412, "customers": 368, "messages": 3104, "agent_replies": 1530,
    "leads": 57, "escalated": 12, "resolved": 140, "human_handled": 9,
    "searches": 820, "unanswered": 64
  },
  "daily": [{ "day": "2026-10-05", "conversations": 18, "replies": 71, "leads": 3 }],
  "hourly": [{ "hour": 9, "conversations": 41 }],
  "channels": [{ "kind": "whatsapp", "conversations": 260 }, { "kind": "web", "conversations": 152 }],
  "knowledgeGaps": [{ "query": "do you deliver on sunday", "asked": 7, "last_asked": "2026-10-05T08:12:00Z", "best_similarity": 0.41 }],
  "topQuestions": [{ "query": "where is my order", "asked": 96, "answered": 94, "last_asked": "2026-10-05T10:02:00Z" }],
  "derived": { "neededAPerson": 21, "answerRate": 0.92 }
}

days is a rolling window, 1–90 (default 30). rows caps the two question lists, 1–100 (default 20).
resolved and escalated count conversations by their current status.
knowledgeGaps are questions the agent searched its knowledge for and found nothing — the list to write new knowledge from.
derived.neededAPerson is escalated plus handled by a person; derived.answerRate is the share of knowledge searches that found an answer (null when there were none).
Days and hours are in UTC.

Call totals

GET /api/v1/calls/summary?days=7

{
  "days": 7,
  "timezone": "Asia/Kolkata",
  "grain": "day",
  "calls": 230, "inbound": 140, "outbound": 90,
  "answered": 198, "missed": 26, "failed": 6,
  "seconds": 41820,
  "debitMicro": 1394000000,
  "surfaces": { "phone": 210, "web": 14, "whatsapp": 6 },
  "series": [{ "bucket": "2026-10-05", "inbound": 22, "outbound": 15, "answered": 33, "missed": 4, "minutes": 118 }]
}

days is 1–90 (default 7). With days=1 the series is by hour. Buckets are in your workspace’s time zone. debitMicro is what the calls cost you.

WhatsApp totals

GET /api/v1/whatsapp/summary?days=7
GET /api/v1/whatsapp/summary?from=2026-09-01&to=2026-09-30

{
  "from": "2026-09-01",
  "to": "2026-09-30",
  "timezone": "Asia/Kolkata",
  "grain": "day",
  "sent": 1820, "delivered": 410, "read": 1290, "failed": 34, "noReceipt": 86,
  "received": 2210,
  "series": [{ "bucket": "2026-09-30", "sent": 61, "received": 70 }],
  "errors": [{ "error": "131049", "n": 21 }],
  "cost": {
    "currency": "INR",
    "totalMicro": 214350000,
    "complete": true,
    "byCategory": [
      { "category": "marketing", "messages": 1200, "costMicro": 191400000 },
      { "category": "utility", "messages": 620, "costMicro": 22950000 }
    ]
  }
}

cost is what the WhatsApp messages in the window cost you, priced exactly as the WhatsApp dashboard prices them, by Meta category for the messages you sent. complete: false means some messages are not priced yet — Meta has not reported them, or there is no rate card for that country — so totalMicro is what is known so far. Either a rolling days (1–90, default 7) or a from/to range (inclusive, up to 366 days), in your workspace’s time zone. errors counts failures by WhatsApp error code — see Not delivered.

Errors

Status	When
`400`	`days`, `rows` or a date range out of bounds
`404`	The agent is not in your workspace

​Usage and balance

​Agent analytics

​Call totals

​WhatsApp totals

​Errors

Usage and balance

Agent analytics

Call totals

WhatsApp totals

Errors