explore-app-store

name

explore-app-store

description

App Store idea discovery and market intelligence for iOS apps, using free public APIs (iTunes Search/Lookup/RSS charts + the no-auth Sensor Tower app endpoint for revenue & download estimates). Use this skill when the user wants to: (1) Find new app ideas or "sweet spots" — recently released apps already earning a target revenue band (e.g. 2026 apps making $10-30k/mo) that look winnable and easy to ship, (2) Map a niche's competitors with revenue/downloads/ratings, (3) Find localization arbitrage — apps big in one country/language but absent or unlocalized in another, including iOS-vs-Android port gaps, (4) Tear down a developer/publisher's portfolio or detect "rising publishers" shipping multiple monetizing apps, (5) Pull top free/paid/grossing charts per storefront and genre, (6) Grab competitor screenshots, App Store links, pricing/paywall, and descriptions into organized folders with a ranked CSV + opportunity brief. Trigger on: "find app ideas", "explore the app store", "trending apps making money", "sweet spot", "competitor research", "what works in <country>", "developer portfolio", "app market research", "localization arbitrage", "apps in <country X> we can apply in <country Y>", "take a local app global".

explore-app-store

Turn a plain-English goal ("find 2026 apps making $10-30k I could clone", "map the hair-loss niche", "what's big in Turkey but missing in the US") into ranked opportunities backed by real revenue/download estimates, screenshots, and App Store links.

All scripts are pure Python stdlib — no dependencies, no API keys. Run them with python3. Every revenue/download number comes from Sensor Tower and is an estimate, coarsely bucketed (<$5k, $10k, $30k, $90k, ...). Present them as relative signal.

Data sources (all public, no auth)

• iTunes Search — keyword discovery
• iTunes Lookup — details by app id / bundleId / developer portfolio (artistId)
• iTunes RSS charts — top free / paid / grossing per country + genre
• Sensor Tower app API — est. worldwide last-month downloads + revenue, publisher, top_countries / valid_countries, screenshots (iOS and Android). Rate-limited (HTTP 429 after ~180 ids) → scripts batch + backoff + cache automatically.

Workflow — how to drive this skill

The scripts do the deterministic fetching/scoring; you (the model) supply the creativity — brainstorm a broad, varied seed-keyword set before sweeping, and interpret the results.

1. Find sweet spots / new ideas → scripts/explore.py sweep

First brainstorm 12-30 seed keywords covering the domain from many angles (synonyms, "AI X", "X tracker/scanner/planner", the problem, the audience). Then:

python3 scripts/explore.py sweep \
  --terms "hair loss,scalp scanner,hair growth,thinning hair,derma roller,minoxidil tracker,..." \
  --release-min 2026 --rev-min 10000 --rev-max 30000 --country us --limit 50 --top 20

Output (in explore-out/<slug>-<ts>/): apps.csv, apps.json, brief.md (ranked with a "why it scores" breakdown). Add --charts to also fold in top-grossing/free chart apps. Adjust --release-min / --rev-min/max to widen or tighten. Read brief.md, then present the top opportunities and your read on which are genuinely winnable.

Sweet-spot score weights: in target revenue band (0.30), recency (0.20), monetization efficiency = revenue÷downloads (0.20), winnability = few ratings (0.15), velocity = ratings÷age (0.10), ship-ability = few screenshots (0.05). See references/scoring.md.

2. Map a niche's competitors → same sweep, no release filter

python3 scripts/explore.py sweep --terms "<niche terms>" --release-min "" --rev-min 0 --top 40

3. Localization arbitrage → scripts/explore.py localize

python3 scripts/explore.py localize --base us --targets tr,de,br,kr,gb,jp \
  --genre 6023 --kind top-grossing --top 20
# and, for specific apps, language coverage + iOS-vs-Android port gaps:
python3 scripts/explore.py localize --ids 6759983589,6466220595 --langs tr,de,es,ja

Reports apps charting in a target market but absent from the base (import candidates, matched by app id so localized names don't create false positives), plus per-app missing languages and whether the iOS app is absent from Google Play (a port opportunity).

3b. Bidirectional arbitrage (both directions at once) → scripts/explore.py arbitrage

Find growing apps to move BETWEEN markets, in both directions, across several genres:

python3 scripts/explore.py arbitrage --home tr --markets us,gb,de \
  --genres 6012,6013,6008,6002,6017 --kind top-grossing --rev-min 10000 --release-min 2025

• EXPORT = apps whose top revenue country is HOME and that are absent from MARKETS → "born in your market, take global" (the high-value direction for a local indie: build once, earn tier-1 ARPU).
• IMPORT = apps charting in MARKETS, absent from HOME, and not localized to HOME's language → "proven abroad, localize in".
• Swap --home/--markets to run the reverse (e.g. --home us --markets tr). Writes arbitrage.json. --home-lang overrides the language derived from --home.

4. Developer / rising-publisher analysis → scripts/explore.py portfolio

python3 scripts/explore.py portfolio --artist 1499186364            # one developer's apps + revenue
python3 scripts/explore.py portfolio --from-json explore-out/<run>/apps.json --min-apps 2

--from-json groups a prior sweep by publisher and flags publishers shipping ≥N recent apps that each monetize — a repeatable playbook to study.

5. Charts → scripts/explore.py charts

python3 scripts/explore.py charts --country us --kind top-grossing,top-free --genre 6007 --limit 50

6. Screenshots + details for a shortlist → scripts/explore.py assets

python3 scripts/explore.py assets --ids 6740919592,6751491972 --max-shots 10 --out research/screenshots

Writes details.md (revenue, paywall pricing, top countries, languages, full description) + screenshots per app.

Typical end-to-end

1. sweep a domain → read brief.md → shortlist the highest sweet-spot scores.
2. portfolio --from-json to spot the studio behind repeated winners.
3. assets the shortlist for screenshots + paywall pricing.
4. localize to check for an untapped market or a missing Android port.
5. Summarize: the opportunity, why it's winnable, the observed monetization model.

Notes & limits

• Revenue is worldwide, last-month, bucketed estimates — relative signal, not exact $. The band filter works on buckets (can't split $12k from $18k).
• top_countries shows what drives revenue; there is no per-country revenue in the free endpoint — localization uses chart presence + language gaps, not exact per-country $.
• Charts surface incumbents (ChatGPT, TikTok), not small indies — sweet spots come from keyword sweeps, so point the sweep at a domain and brainstorm seeds generously.
• Be honest that these are estimates; don't overstate precision.
• References: references/genre_ids.md, references/storefronts.md, references/api_notes.md, references/scoring.md.