Before I build an app, I want to know three things: who already makes money in the niche, what people search for, and what users complain about in competitors' reviews. That used to mean a dozen browser tabs and a spreadsheet. With the Model Context Protocol (MCP), Claude can call those lookups as tools in the middle of a conversation.
Disclosure: I build GetAppNiche, the data source in this walkthrough. It's a paid product ($39/month); the MCP server and REST API come with the plan.
What the server exposes
GetAppNiche runs a hosted MCP server at https://api.getappniche.com/mcp (Streamable HTTP, bearer-token auth). It has eight read-only tools:
-
search_apps— search App Store and Google Play apps with 25+ filters (store, category, rating, review count, estimated revenue and downloads, release date, review growth) -
get_app_detail— one app's full record -
get_app_historicals— metric history for one app -
get_app_timeline— what changed in an app's listing over time -
get_keyword_difficultyandbatch_keyword_difficulty— keyword research on either store -
get_app_reviews— live store reviews for any app get_supported_countries
Revenue and download figures are estimates from GetAppNiche's own model, and the tools label them that way. Keep that in mind when you read the answers.
Connect it
Claude Code — one command:
claude mcp add --transport http getappniche https://api.getappniche.com/mcp --header "Authorization: Bearer YOUR_API_KEY"
Start a session and ask "What GetAppNiche tools do you have?" to check it worked. A 401 means the key didn't make it into the header.
Claude Desktop — add the official bridge to claude_desktop_config.json (Node 18+):
{
"mcpServers": {
"getappniche": {
"command": "npx",
"args": ["-y", "@getappniche/mcp"],
"env": { "GETAPPNICHE_API_KEY": "YOUR_API_KEY" }
}
}
}
Cursor and other clients: the snippets are on getappniche.com/mcp.
Three prompts I actually use
1. Is there money in this niche?
Find habit tracker apps on the App Store with fewer than 5,000 reviews that earn an estimated $5k+ a month.
Claude calls search_apps with a search term, the store, a review-count cap and a revenue floor, then summarizes. (Revenue estimates are strongest on the App Store; on Google Play I filter by downloads and reviews instead.) Small apps with real estimated revenue are the signal I look for; a niche where only giants earn is a different bet.
2. How does a competitor compare across stores?
Pull the app details for Calm on both stores and compare estimated downloads, ratings and review counts.
Two get_app_detail calls and a side-by-side table. Useful for seeing which store an app really lives on.
3. What do users hate?
Read the latest 100 reviews of [competitor] and list the five most common complaints, with one quote each.
get_app_reviews reads live store reviews for any app, so this works for apps you don't track. The complaint list often turns into a feature list.
Limits worth knowing
- Every call spends API credits; the Pro plan includes 5,000 a month. List results cost per row, so ask for 20 apps, not 500.
- The tools are read-only. Daily keyword rank tracking lives in the GetAppNiche app, not in MCP.
- 14–90-day growth windows and trend history are App Store only today.
If you'd rather try the data without an API key, the free keyword tool shows who ranks for any keyword on either store, and the revenue checker looks up any iOS app. Both work without signing up.
Top comments (0)