Ask the App Store anything.

Name: app-store-mcp
Author: Perttu Lähteenlahti

app-store-mcp gives Claude and ChatGPT live access to Apple's App Store. Search apps, read reviews, follow charts. No API key.

Get started View on GitHub

Claude app-store-mcp connected

You

Which meditation apps have big audiences but unhappy users?

app_store_search app_store_ratings app_store_reviews

Claude

Three of the top ten have six-figure rating counts but average under 3.6 stars. Their one-star reviews agree on the reason: people love the content and hate the paywall.

An example conversation. The tool calls are real.

Every public listing, chart, and review in the catalog. These icons are today's top free apps.

Set up in under a minute.

No account, no keys, no build step. Pick your client below; the server downloads and runs itself the first time your assistant calls it.

Run this in a terminal, then type /mcp inside Claude Code to confirm the server shows up.

claude mcp add app-store -- npx -y @perttu/app-store-mcp

Open Settings › Developer › Edit Config, add this to claude_desktop_config.json, and restart Claude.

{
  "mcpServers": {
    "app-store": {
      "command": "npx",
      "args": ["-y", "@perttu/app-store-mcp"]
    }
  }
}

Add this to ~/.cursor/mcp.json, or paste it under Settings › MCP › Add new server.

{
  "mcpServers": {
    "app-store": {
      "command": "npx",
      "args": ["-y", "@perttu/app-store-mcp"]
    }
  }
}

ChatGPT reaches connectors over the web, so the server needs a public address. Start a local bridge first:

npx -y supergateway --stdio "npx -y @perttu/app-store-mcp" --port 8686

Then expose port 8686 with a tunnel such as ngrok, and add the public /sse URL as a connector under Settings › Connectors › Advanced › Developer mode.

Any MCP client that can launch a local stdio server works the same way, including VS Code, Zed, and Windsurf. You need Node 20 or newer.

Research that used to take an afternoon.

You don't write API calls. You ask a question, and your assistant picks the right tools, makes the requests, and reads everything so you don't have to.

Find the gap in a market

Look at the top 50 habit trackers. Which ones are rated under four stars, and what do their one-star reviews complain about?

Six apps with six-figure audiences average under 3.8
Lost three years of streaks after the update
Widgets stopped syncing and support never answered
Nobody has a working offline mode

app_store_list app_store_ratings app_store_reviews

Watch a competitor

What has Duolingo shipped in its last five updates, and how did users react?

app_store_version_history app_store_reviews

Check the pricing

What do the top grossing meditation apps actually charge for subscriptions?

app_store_list app_store_in_app_purchases

Mine the search box

What does the store suggest after “sleep”, and who ranks first for each phrase?

app_store_suggest app_store_search

Read the privacy label

What data does CapCut say it collects, and is any of it linked to identity?

app_store_privacy

Somewhere in a million one‑star reviews is the app someone wishes existed. Now you can ask.

From category to app idea in four questions.

The loop indie developers run with it: start wide, find the apps users feel stuck with, then read exactly why they stay unhappy.

Get started

01Map the field

One question returns the whole chart, with names, ratings, and rating counts attached.

List the top 100 free Health & Fitness apps with their ratings.

app_store_list

02Find the stuck users

A big audience with a low average means demand is proven and the current options disappoint.

Which of those have over 10,000 ratings but average under 3.8?

No new requests here. The model filters what it already fetched.

03Read the pain

Complaints cluster fast. A few pages in, you know the top three failures by heart.

Pull recent one-star reviews for the worst three. What keeps coming up?

app_store_reviewsapp_store_ratings

04Build what's missing

You're not guessing anymore. You're answering feature requests people already wrote down, in public, with dates on them.

Eleven tools. One server. No keys.

Everything is read-only and comes back as structured JSON alongside formatted text, so any MCP client gets a response it can use.

app_store_app

The full listing for one app, by track ID or bundle ID.

app_store_search

Keyword search with pagination, or IDs only for follow-up calls.

app_store_list

Top free, paid, and grossing charts, filterable by category.

app_store_reviews

Recent or most helpful reviews, up to ten pages deep.

app_store_ratings

Total count and the full one-to-five star distribution.

app_store_developer

Every app published by a single developer.

app_store_similar

The apps the store lists alongside a given one.

app_store_suggest

Autocomplete suggestions for a search term.

app_store_privacy

Privacy labels: what an app collects and how it's linked.

app_store_version_history

Past versions with dates and release notes.

app_store_in_app_purchases

Top in-app purchase names with localized prices.

Every storefront

Add a two-letter country code like fi, gb, or jp to any call. Defaults to the US store.

Common questions.

Do I need an Apple developer account or an API key?

No. The server reads the same public endpoints the App Store website uses. Install it and start asking questions.

Does it work with ChatGPT?

Yes. ChatGPT's developer-mode connectors talk to MCP servers over HTTP, so you expose the server through a small bridge with a public URL. The ChatGPT tab above walks through it.

Is the data live?

Yes. Every answer comes from a fresh request at the moment you ask. Nothing is cached, so ranks, ratings, and reviews are current.

What does it cost?

Nothing. It's MIT-licensed open source. Apple's endpoints are public, though Apple can rate-limit or change them without notice.

Which countries does it cover?

Any storefront. Pass a two-letter country code such as fi, gb, or jp to any tool. It defaults to the US store.

Is this affiliated with Apple?

No. It's an independent open-source project. App Store is a trademark of Apple Inc.