App Store Autocomplete as a Keyword Research Tool

The best keyword source for ASO isn't a paid database — it's the store's own search bar. Autocomplete suggestions are real queries typed by real users, kept current by the store itself, and returned in rough popularity order. ASORadar's keyword-ideas endpoint mines that source programmatically, for any country, on both stores.

Why autocomplete beats guessing

When you (or an LLM) invent keywords, you get phrases that sound right. Autocomplete only contains phrases people actually search — the store won't suggest what nobody types. That single property removes the biggest failure mode of DIY keyword research: optimizing for queries with zero traffic.

Mining the suggestion tree with one call

curl "https://api.asoradar.com/a1/keyword-ideas?term=budget&country=us&access_key=YOUR_KEY"

{
  "ok": true,
  "data": {
    "term": "budget",
    "country": "us",
    "expanded": false,
    "ideas": [
      { "keyword": "budget app",        "sources": ["base"] },
      { "keyword": "budget planner",    "sources": ["base"] },
      { "keyword": "budgeting apps free", "sources": ["base"] }
    ],
    "count": 10
  }
}

A single call returns what the search bar would suggest for your seed. Add expand=true and it walks the whole tree — budget a, budget b, … budget z (27 upstream calls, billed accordingly) — then dedupes everything into one list. That surfaces the long-tail phrases a human would never think to type letter by letter. /g1/keyword-ideas does the same for Google Play (add lang).

Reading demand honestly

Two signals come free with every idea. Position: autocomplete is popularity-ordered, so earlier suggestions are searched more. Sources: with expand=true, each idea lists which prefixes surfaced it — a phrase that shows up under base AND half the alphabet is in demand across the tree, not a one-prefix fluke.

These are ordinal signals, not volume numbers — and that's deliberate. ASORadar doesn't fabricate search-volume scores; if you need absolute keyword popularity, pair this with Apple Search Ads data. For choosing between candidate keywords, order + sources + a rank check is usually all an indie needs.

From ideas to positions

Found candidates? Check where you (and who) stand on each with keyword-rank — live position plus the top-3 apps owning each phrase, up to 10 terms per call. Then subscribe a keyword_rank webhook and get a signed POST when your position moves.

Do it per country — never translate keywords

Autocomplete is per-storefront: pass country=de, country=br, country=in and you get what people in that market actually type. This is the honest replacement for the classic DIY mistake — machine-translating your US keyword list. Translated keywords often have zero local search traffic, and some markets don't search in the local language at all (India searches mostly in English, adapted to local intent). Mine each market's own autocomplete instead of translating another market's list.

Pricing

Pay per request, from $0.001: one call without expand, 27 with it. No subscription, no keyword caps. 100 free requests to try — no card. Full parameter reference in the API docs; also available as an MCP tool so your AI agent can run keyword research end-to-end.