# Get aggregate outbound Apple Messages for Business statistics

`GET /v1/amb/stats/summary`

Returns one aggregate row of outbound Apple Messages for Business counts and latency percentiles (`p50`, `p95`, and `p99`) for the requested period. Counts are attributed to the time each message was accepted, so a send failure recorded today counts against the day its message was accepted; `period.data_as_of` is the freshness boundary the rollups have reached.

There is no delivery receipt on this channel, so there is no `delivered` count, delivery rate, or delivery-latency figure anywhere in this response. `latency.processing` covers acceptance to Apple handoff, and `first_response` measures the oldest unanswered inbound message to acceptance of the first outbound reply that claims it. When no qualifying message contributes a measurement, processing percentiles are null and `first_response` is omitted.

`from` and `to` must both be calendar days or RFC 3339 instants. Day windows cover up to 365 whole days. Instant bounds round down to the hour, remain inclusive, and may span up to 720 hours. Mixing the forms returns `422`. Set `timezone` for local boundaries. Historical dates are supported within the maximum window length; the requested dates are not shifted forward.

Set one dimension filter at most (`business_id`, `message_kind`, `intent`, `group`, `category`, or `tag`); more than one returns `422`. Use `compare=previous_period` to include the preceding equal-length window and each metric's change.

## Code samples

**TypeScript**

```ts
const result = await bird.amb.stats.summary({
  business_id: "abz_01krdgeqcxet5s7t44vh8rt9mg",
});
console.log(result);
```

Examples: [TypeScript](/docs/api/reference/get-amb-stats-summary.ts.md) · [Python](/docs/api/reference/get-amb-stats-summary.py.md) · [Go](/docs/api/reference/get-amb-stats-summary.go.md) · [PHP](/docs/api/reference/get-amb-stats-summary.php.md) · [CLI](/docs/api/reference/get-amb-stats-summary.cli.md) · [MCP](/docs/api/reference/get-amb-stats-summary.mcp.md) · [cURL](/docs/api/reference/get-amb-stats-summary.curl.md)

## Example response `200`

```json
{
  "period": {
    "from": "2026-05-01",
    "to": "2026-05-25",
    "data_as_of": "2026-05-25T14:03:10Z"
  },
  "attribution": "accepted_time",
  "counts": {
    "accepted": 4820,
    "sent": 4790,
    "send_failed": 30,
    "rejected": 4,
    "sent_rate": 0.9938,
    "send_failure_rate": 0.0062
  },
  "latency": {
    "processing": {
      "p50_ms": 610,
      "p95_ms": 2140,
      "p99_ms": 5380
    }
  },
  "first_response": {
    "p50_ms": 610,
    "p95_ms": 2140,
    "p99_ms": 5380
  },
  "comparison": {
    "period": {
      "from": "2026-05-01",
      "to": "2026-05-25",
      "data_as_of": "2026-05-25T14:03:10Z"
    },
    "counts": {
      "accepted": 4820,
      "sent": 4790,
      "send_failed": 30,
      "rejected": 4,
      "sent_rate": 0.9938,
      "send_failure_rate": 0.0062
    },
    "latency": {
      "processing": {
        "p50_ms": 610,
        "p95_ms": 2140,
        "p99_ms": 5380
      }
    },
    "first_response": {
      "p50_ms": 610,
      "p95_ms": 2140,
      "p99_ms": 5380
    },
    "delta": {
      "accepted_pct_change": 0.508,
      "sent_pct_change": 0.501,
      "send_failed_pct_change": -0.12,
      "rejected_pct_change": 0
    }
  }
}
```

## Query parameters

- `from` (string): Inclusive start of the window: a calendar day (YYYY-MM-DD) or an RFC 3339 instant rounded down to the hour. The `timezone` parameter makes a calendar day local and rounds an instant down to the local hour. Omit `timezone` to use UTC. When `timezone` is set, a numeric UTC offset such as `+05:45` is rejected; use a calendar day or a `Z` (UTC) instant. This value must use the same form as `to`. When omitted, it defaults to 30 days before `to` for day windows or 168 hours (7 days) before `to` for hour windows.
- `to` (string): Inclusive end of the window: a calendar day (YYYY-MM-DD) or an RFC 3339 instant rounded down to the hour. The `timezone` parameter makes a calendar day local and rounds an instant down to the local hour. Omit `timezone` to use UTC. When `timezone` is set, a numeric UTC offset is rejected; use a calendar day or a `Z` (UTC) instant. This value must use the same form as `from`. When omitted, it defaults to today for day windows or the current hour for hour windows in that timezone. Day windows may not exceed 365 days; hour windows may not exceed 720 hours (30 days).
- `timezone` (string): IANA timezone identifier used to group statistics, for example `Asia/Kathmandu`. The default is UTC. Day and hour boundaries, including the default window when `from` and `to` are omitted, follow this timezone. When this parameter is set, pass `from` and `to` as calendar days or `Z` instants instead of timestamps with explicit UTC offsets.
- `business_id` (string): Restrict the statistics to a single business. Mutually exclusive with the other dimension filters (`message_kind`, `intent`, `group`, `category`, `tag`); only one may be set per request. Matches the `business_id` key on a row of the per-business breakdown.
- `message_kind` (string)

  Restrict the statistics to a single message content kind. Mutually exclusive with the other dimension filters (`business_id`, `intent`, `group`, `category`, `tag`); only one may be set per request. Matches the `message_kind` key on a row of the per-content-kind breakdown.

  Possible values: `text`, `attachment`, `rich_link`, `quick_reply`, `list_picker`, `time_picker`, `form`, `apple_pay`, `authenticate`, `imessage_app`, `interactive`
- `intent` (string): Restrict the statistics to a single entry-point intent. Mutually exclusive with the other dimension filters (`business_id`, `message_kind`, `group`, `category`, `tag`); only one may be set per request. Matches the `intent` key on a row of the per-intent breakdown.
- `group` (string): Restrict the statistics to a single entry-point group. Mutually exclusive with the other dimension filters (`business_id`, `message_kind`, `intent`, `category`, `tag`); only one may be set per request. Matches the `group` key on a row of the per-group breakdown.
- `category` (string): Restrict the statistics to a single message category. Mutually exclusive with the other dimension filters (`business_id`, `message_kind`, `intent`, `group`, `tag`); only one may be set per request. Matches the `category` key on a row of the per-category breakdown.
- `tag` (string): Restrict the statistics to a single tag. Use `name` to match any value of a tag, or `name:value` for a specific pair (for example `campaign:spring_launch`). Mutually exclusive with the other dimension filters (`business_id`, `message_kind`, `intent`, `group`, `category`); only one may be set per request. A row of the per-tag breakdown carries the same pair in its single `tag` key.
- `compare` (string)

  Set to `previous_period` to also include the same statistics for the immediately preceding window of equal length, plus the change between the two, so you can show "+X% vs last period" without a second request. The comparison window carries any dimension filter set on the request, so a filtered comparison compares like with like.

  Possible values: `previous_period`

## Response body

- `period` (object, required): The window the response covers (echoed back from the request), plus `data_as_of`, the freshness boundary the data is current to.
- `period.from` (string, required): Inclusive start of the window, as a calendar day (`YYYY-MM-DD`) or an RFC 3339 hour boundary. Historical starts are preserved; the maximum request length does not impose a historical cutoff.
- `period.to` (string, required): Inclusive end of the window, as a calendar day (`YYYY-MM-DD`) or an RFC 3339 hour boundary.
- `period.data_as_of` (nullable string): Latest time reflected in the statistics. More recent events might not be included yet. Null when the freshness boundary is unavailable.
- `attribution` (string, required)

  Always `accepted_time` for outbound statistics.

  Possible values: `accepted_time`, `event_time`
- `counts` (object, required): Outbound Apple Messages for Business counts for the requested scope, attributed to when each message was accepted. Apple Messages for Business has no delivery receipt, so there is no `delivered` count anywhere in this API: `sent` is the last outbound state Bird observes for a message. Very large counts are close estimates rather than exact tallies. Rates are computed once here, clamped to 1, and null when nothing was accepted.
- `counts.accepted` (integer, required): Distinct messages accepted for sending after admission checks. This is the denominator for `sent_rate` and `send_failure_rate`.
- `counts.sent` (integer, required): Distinct messages handed off to Apple.
- `counts.send_failed` (integer, required): Distinct accepted messages that Apple refused or that exhausted their send attempts. See `last_error.code` on the message for the reason; a refused charge is not a send failure, it is `rejected`.
- `counts.rejected` (integer, required): Distinct messages refused before any send attempt, because the destination has no price, the wallet could not fund the send, or the content cannot be sent yet. Rejected messages are never charged and are not counted in `accepted`, so the total addressed is `accepted + rejected`. Excluded from `send_failure_rate`, which covers send failures only.
- `counts.sent_rate` (nullable number, required): Share of accepted messages Apple acknowledged, computed as `sent / accepted`. Null when no messages were accepted in scope. This stands where other channels report a delivery rate.
- `counts.send_failure_rate` (nullable number, required): Share of accepted messages that failed to send, computed as `send_failed / accepted`. Null when no messages were accepted in scope.
- `latency` (object, required): Processing-latency percentiles in milliseconds for the requested scope, from acceptance to Apple handoff. Apple Messages for Business has no delivery receipt, so there is no `delivery` or `total` member beside `processing`. Conversation response timing is reported separately in `first_response`. Always present; every percentile is null when no qualifying message in scope has a measurement.
- `latency.processing` (object, required): Approximate p50, p95, and p99 latency percentiles in milliseconds for one latency family. All three are null when no qualifying event contributed a measurement.
- `latency.processing.p50_ms` (nullable integer, required): Median (50th percentile) latency in milliseconds. Null when no qualifying event contributed a measurement.
- `latency.processing.p95_ms` (nullable integer, required): 95th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `latency.processing.p99_ms` (nullable integer, required): 99th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `first_response` (object): Time from the oldest unanswered inbound message to acceptance of the first outbound reply that claims it. Reported on summaries and breakdowns. Omitted when no qualifying reply contributes a measurement.
- `first_response.p50_ms` (nullable integer, required): Median (50th percentile) latency in milliseconds. Null when no qualifying event contributed a measurement.
- `first_response.p95_ms` (nullable integer, required): 95th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `first_response.p99_ms` (nullable integer, required): 99th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison` (object): The same statistics for the equal-length, inclusive period ending immediately before the requested start, together with the change between the two periods. Present only when `compare=previous_period` is requested. The change is already computed, so a percentage difference needs no second request.
- `comparison.period` (object, required): The preceding window these comparison figures cover, the equal-length window ending immediately before the requested start (the prior day for day windows, the prior hour for hour windows).
- `comparison.period.from` (string, required): Inclusive start of the window, as a calendar day (`YYYY-MM-DD`) or an RFC 3339 hour boundary. Historical starts are preserved; the maximum request length does not impose a historical cutoff.
- `comparison.period.to` (string, required): Inclusive end of the window, as a calendar day (`YYYY-MM-DD`) or an RFC 3339 hour boundary.
- `comparison.period.data_as_of` (nullable string): Latest time reflected in the statistics. More recent events might not be included yet. Null when the freshness boundary is unavailable.
- `comparison.counts` (object, required): Outbound Apple Messages for Business counts for the requested scope, attributed to when each message was accepted. Apple Messages for Business has no delivery receipt, so there is no `delivered` count anywhere in this API: `sent` is the last outbound state Bird observes for a message. Very large counts are close estimates rather than exact tallies. Rates are computed once here, clamped to 1, and null when nothing was accepted.
- `comparison.counts.accepted` (integer, required): Distinct messages accepted for sending after admission checks. This is the denominator for `sent_rate` and `send_failure_rate`.
- `comparison.counts.sent` (integer, required): Distinct messages handed off to Apple.
- `comparison.counts.send_failed` (integer, required): Distinct accepted messages that Apple refused or that exhausted their send attempts. See `last_error.code` on the message for the reason; a refused charge is not a send failure, it is `rejected`.
- `comparison.counts.rejected` (integer, required): Distinct messages refused before any send attempt, because the destination has no price, the wallet could not fund the send, or the content cannot be sent yet. Rejected messages are never charged and are not counted in `accepted`, so the total addressed is `accepted + rejected`. Excluded from `send_failure_rate`, which covers send failures only.
- `comparison.counts.sent_rate` (nullable number, required): Share of accepted messages Apple acknowledged, computed as `sent / accepted`. Null when no messages were accepted in scope. This stands where other channels report a delivery rate.
- `comparison.counts.send_failure_rate` (nullable number, required): Share of accepted messages that failed to send, computed as `send_failed / accepted`. Null when no messages were accepted in scope.
- `comparison.latency` (object, required): Processing-latency percentiles in milliseconds for the requested scope, from acceptance to Apple handoff. Apple Messages for Business has no delivery receipt, so there is no `delivery` or `total` member beside `processing`. Conversation response timing is reported separately in `first_response`. Always present; every percentile is null when no qualifying message in scope has a measurement.
- `comparison.latency.processing` (object, required): Approximate p50, p95, and p99 latency percentiles in milliseconds for one latency family. All three are null when no qualifying event contributed a measurement.
- `comparison.latency.processing.p50_ms` (nullable integer, required): Median (50th percentile) latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison.latency.processing.p95_ms` (nullable integer, required): 95th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison.latency.processing.p99_ms` (nullable integer, required): 99th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison.first_response` (object): First-response latency percentiles for the preceding period. Omitted when no qualifying message in that period has a measurement.
- `comparison.first_response.p50_ms` (nullable integer, required): Median (50th percentile) latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison.first_response.p95_ms` (nullable integer, required): 95th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison.first_response.p99_ms` (nullable integer, required): 99th percentile latency in milliseconds. Null when no qualifying event contributed a measurement.
- `comparison.delta` (object, required): Changes from the previous period. Each value is the signed relative change `(current - previous) / previous` and is null when the previous count is zero.
- `comparison.delta.accepted_pct_change` (nullable number, required): Relative change in accepted messages (`counts.accepted`) versus the previous period, as a signed fraction. Null when the previous period accepted none.
- `comparison.delta.sent_pct_change` (nullable number, required): Relative change in sent messages (`counts.sent`) versus the previous period, as a signed fraction. Null when the previous period had none.
- `comparison.delta.send_failed_pct_change` (nullable number, required): Relative change in send failures (`counts.send_failed`) versus the previous period, as a signed fraction. Null when the previous period had none.
- `comparison.delta.rejected_pct_change` (nullable number, required): Relative change in rejected messages (`counts.rejected`) versus the previous period, as a signed fraction. Null when the previous period had none.

## Related resources

- [Should I use a Bird SDK or call the API directly?](/explained/platform/should-i-use-an-sdk-or-call-the-api-directly) (answer)
- [Build your first integration](/learn/paths/integration) (course)
- [Send your first email](/docs/get-started/send-your-first-email) (docs)

[Get an implementation brief](/learn/workspace?topic=api-basics)