Method and limits

What this is

For each market (Korea, US, Japan App Store), it uses public data to show how crowded a keyword space is and what users of competing apps complain about, so a solo developer can judge whether there is a gap worth building.

What this is not

• ·Not a search-volume ranking. Apple does not publish App Store search volume.
• ·Not a success predictor. Most indie apps fail even with a good idea.
• ·Not an ASO optimization tool. Tools for teams that already have an app exist elsewhere.

How keywords are chosen

Not hand-picked. For each market and category, an LLM (subscription batch) generates candidate search terms a real user might type, then every candidate is run against the iTunes Search API and only those with results are kept. The keyword list is LLM-generated; the per-keyword data is 100% real API output.

Data sources

• ·iOS (official, free): iTunes Search API for search results and competitor metadata, the public App Store review feed for recent reviews. A public endpoint Apple recommends caching for heavy use.
• ·Android (unofficial): Google Play has no official free API. We scrape the public web store with google-play-scraper. That is a ToS gray area and can break when Google changes its HTML, so coverage may be less stable than iOS. We label it as not official data.

How the saturation signal is computed

Not a black box. It uses the review concentration of the top 10 apps, not the result count. A #1 app with 10k+ reviews, or a top-10 sum over 30k, means red ocean; below thresholds is mid; only small apps means an open gap. In a direct test, photo cleaner and a niche term had similar result counts but top-10 review sums of 45,000 vs 413.

Honesty guards

• ·If there are fewer than 5 review samples, we leave complaints blank instead of inventing them.
• ·If there are too few result apps, we mark it low signal.
• ·Every review excerpt is an actual sentence from a real review.

The biggest limits (please read)

• ·Search demand is not success. Apps are usually discovered through ads, social, and word of mouth more than search. This tool helps with part of 'what to build', not 'how to get it found'. The real bottleneck is usually distribution.
• ·The review feed exposes at most the latest 500 per app, so it is strong on recent complaint patterns but not a full historical analysis.
• ·Similar tools already exist (GapFind, PainMap and others). Globally we compete with them; local-language phrasing and context are where we go deeper.

Cost and privacy

• ·Heavy work like review analysis is pre-computed in a local batch and stored statically. We do not call an AI in real time when you open a page.
• ·No tracking scripts.