"ai chat"), returns the per-day record of every prefix that surfaced it in Apple’s autocomplete, plus the best rank Apple placed it in. Use this to confirm whether a suggestion is sticky or a one-day flicker, and to understand which prefix paths users follow to discover it.
Query Parameters
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| suggestion | string | Yes | — | Exact autocomplete suggestion |
| country | string | No | us | ISO 3166-1 alpha-2 storefront |
| days | number | No | 30 | Lookback window in days, 3–180 |
Response
isCurrentlyVisible is true when the suggestion appeared in today’s snapshot — a quick way to confirm whether the trend is still live before acting on it.
trend
Window-over-window read of how Apple’s perception of the suggestion is moving:
| Field | Type | Meaning |
|---|---|---|
| direction | string | promoted (rank improved by ≥ 2), demoted (worsened by ≥ 2), stable, fresh (only one day of data), unknown (no observations). |
| startRank | number | null | Best rank on the first day of the window. |
| endRank | number | null | Best rank on the last observed day. |
| rankImprovement | number | null | startRank − endRank — positive = improved (lower rank number = higher in dropdown). |
| newPrefixes | string[] | Prefixes seen in the most recent 7 days that weren’t seen earlier in the window — Apple expanding the suggestion into new prefix neighbourhoods is the strongest possible “demand rising” signal we can see. |
interpretation is a one-line plain-English summary of the trend — safe to surface directly in a UI and useful for LLM agents.
The response-level meta envelope describes data freshness — see Keyword Metrics → meta envelope for the schema.
Data freshness
The suggestion archive is refreshed daily per storefront, so the response reflects the most recent overnight snapshot. Use the response-levelmeta envelope above to read the exact lastScrapedAt and freshness band for each call.
Credit Cost
2 credits per request.Use Cases
- Decide whether to invest in ranking for a suggestion based on its history (sticky vs one-off).
- Detect when a previously-strong suggestion stops being surfaced (signal of changing search behaviour).
- Build a “discovery path” view: which prefixes users type to reach the suggestion.

