╔═══╗╔══╗╔══╗  ╦═══╦╔═══╦╦  ╦╦╔═══╦╦╗  ╔╦╗╔═══╦╔══╗
║╔═╗║╠╣  ║╠╣    ╠╦╗ ║║╔══╝║  ║║╠╦══╝║║╚╗ ║║║║╔═╗║╠╣
║╚═╝║╚╦╝  ║║    ║╚╝ ║║╚══╗╚╗╔╝║║║╔══╝║║ ╚╗║║║║║ ╠╣╚╦╝
╠══╗║ ║║  ║╚══╗╠╦╝ ║║╔══╗ ╚╦╝ ║║║╚══╗║║  ╚╬╝║║╚═╝║ ║║
╚═══╝╚══╝╚════╝╚╝  ╚╝╚═══╝  ╩  ╩╩╚═══╩╩╝   ╚╝╚════╝╚══╝

// model context protocol · app store reviews · 200ms p95

appreviewmcp — node

mcp@appreviewmcp ~ get_reviews({ app: "com.spotify.music", limit: 5, country: "us" })

// install + run in one command — no config needed for local dev

terminal

$ bunx @appreviewmcp/server

>Get an API key $GitHub

5M+ reviews · 170 countries · 100 req/mo free · no credit card

operational

│

uptime:99.97%·

p95:187ms·

req/d:2.1M·

reviews indexed:5M+·

countries:170+

setup

Up in three commands.

// install — runs without a global install via bunx / npx

terminal

$ bunx @appreviewmcp/server

# or with npm
$ npx -y @appreviewmcp/server

// configure — add to your claude_desktop_config.json

~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "appreviewmcp": {
      "command": "npx",
      "args": ["-y", "@appreviewmcp/server"],
      "env": {
        "API_KEY": "your_key_here"
      }
    }
  }
}

// run — ask Claude (or Cursor) directly in natural language

Claude Desktop

> What are the top complaints about Spotify on iOS
  in the US this month?

// Claude will call get_reviews() and summarise for you

api

REST endpoints. MCP tools. Same data.

// use the REST API directly, or let your AI call the MCP tools automatically

GET/reviews

// fetch reviews for any app — filter by rating, country, date

curl "https://api.appreviewmcp.com/reviews?app=com.spotify.music&country=us&rating=1&limit=20" \
  -H "Authorization: Bearer $API_KEY"

GET/apps/:id/sentiment

// aggregate sentiment score — positive / neutral / negative breakdown

curl "https://api.appreviewmcp.com/apps/com.spotify.music/sentiment?period=30d" \
  -H "Authorization: Bearer $API_KEY"

GET/apps/search

// search apps by name across App Store + Google Play

curl "https://api.appreviewmcp.com/apps/search?q=spotify" \
  -H "Authorization: Bearer $API_KEY"

GET/apps/:id/versions

// compare review trends across app versions — spot regressions

curl "https://api.appreviewmcp.com/apps/com.spotify.music/versions" \
  -H "Authorization: Bearer $API_KEY"

GET/countries

// list all supported country codes — 170+ markets

curl "https://api.appreviewmcp.com/countries" \
  -H "Authorization: Bearer $API_KEY"

// full API reference → /api/docs

compatibility

Works with every major MCP client.

// standard MCP protocol — if it supports MCP, it supports AppReviewMCP

Claude Desktop[OK]

claude_desktop_config.json

command: "npx", args: ["-y", "@appreviewmcp/server"]

Claude Code[OK]

.mcp.json

command: "bunx @appreviewmcp/server"

Cursor[OK]

.cursor/mcp.json

command: "bunx @appreviewmcp/server"

Windsurf[OK]

~/.codeium/windsurf/mcp_config.json

command: "npx", args: ["-y", "@appreviewmcp/server"]

Continue[OK]

~/.continue/config.json

"transport": "stdio"

// mcpServers array

Zed[OK]

~/.config/zed/settings.json

"command": "bunx @appreviewmcp/server"

// context_servers key

// missing your client? open an issue on GitHub

from the changelog

What devs actually shipped.

commita3f9b1cfeat: add review sentiment to product-health dashboard
“Wired AppReviewMCP into my Claude workflow last week. Now I ask 'what broke in the last update?' and get actual user signal back in seconds. Replaced a 3-hour manual process.”
—@priyabuilds·Senior iOS engineer, Bangalore

commitd72e4a8chore: replace manual review scraping with MCP tool call
“The curl → Claude pipeline is dead simple. Point it at your app ID and you get structured review data in whatever format your LLM needs. No scraping, no rate-limit nightmares.”
—@marcuswrites·Full-stack dev, Berlin

commitc19b5f3docs: add MCP-based review analysis to ADR-042
“Used this for competitive analysis. Queried 15 competitor apps across 6 countries in one Cursor session. The 170-country coverage is the real differentiator — nothing else has it.”
—@elenadev·Product engineer, Warsaw

get started

$ bunx @appreviewmcp/server

✓ 100 requests / month — free, forever

✓ no credit card required

✓ 5M+ reviews · 170 countries · 200ms p95

✓ works with Claude, Cursor, Windsurf, Continue, Zed

Get a free API key →Star on GitHub

╔═══╗╔══╗╔══╗ ╦═══╦╔═══╦╦ ╦╦╔═══╦╦╗ ╔╦╗╔═══╦╔══╗ ║╔═╗║╠╣ ║╠╣ ╠╦╗ ║║╔══╝║ ║║╠╦══╝║║╚╗ ║║║║╔═╗║╠╣ ║╚═╝║╚╦╝ ║║ ║╚╝ ║║╚══╗╚╗╔╝║║║╔══╝║║ ╚╗║║║║║ ╠╣╚╦╝ ╠══╗║ ║║ ║╚══╗╠╦╝ ║║╔══╗ ╚╦╝ ║║║╚══╗║║ ╚╬╝║║╚═╝║ ║║ ╚═══╝╚══╝╚════╝╚╝ ╚╝╚═══╝ ╩ ╩╩╚═══╩╩╝ ╚╝╚════╝╚══╝

Up in three commands.

// install — runs without a global install via bunx / npx

terminal

$ bunx @appreviewmcp/server

# or with npm
$ npx -y @appreviewmcp/server

// configure — add to your claude_desktop_config.json

~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "appreviewmcp": {
      "command": "npx",
      "args": ["-y", "@appreviewmcp/server"],
      "env": {
        "API_KEY": "your_key_here"
      }
    }
  }
}

// run — ask Claude (or Cursor) directly in natural language

Claude Desktop

> What are the top complaints about Spotify on iOS
  in the US this month?

// Claude will call get_reviews() and summarise for you

REST endpoints. MCP tools. Same data.

// use the REST API directly, or let your AI call the MCP tools automatically

GET/reviews

// fetch reviews for any app — filter by rating, country, date

curl "https://api.appreviewmcp.com/reviews?app=com.spotify.music&country=us&rating=1&limit=20" \
  -H "Authorization: Bearer $API_KEY"

GET/apps/:id/sentiment

// aggregate sentiment score — positive / neutral / negative breakdown

curl "https://api.appreviewmcp.com/apps/com.spotify.music/sentiment?period=30d" \
  -H "Authorization: Bearer $API_KEY"

GET/apps/search

// search apps by name across App Store + Google Play

curl "https://api.appreviewmcp.com/apps/search?q=spotify" \
  -H "Authorization: Bearer $API_KEY"

GET/apps/:id/versions

// compare review trends across app versions — spot regressions

curl "https://api.appreviewmcp.com/apps/com.spotify.music/versions" \
  -H "Authorization: Bearer $API_KEY"

GET/countries

// list all supported country codes — 170+ markets

curl "https://api.appreviewmcp.com/countries" \
  -H "Authorization: Bearer $API_KEY"

// full API reference → /api/docs

Works with every major MCP client.

// standard MCP protocol — if it supports MCP, it supports AppReviewMCP

Claude Desktop[OK]

claude_desktop_config.json

command: "npx", args: ["-y", "@appreviewmcp/server"]

Claude Code[OK]

.mcp.json

command: "bunx @appreviewmcp/server"

Cursor[OK]

.cursor/mcp.json

command: "bunx @appreviewmcp/server"

Windsurf[OK]

~/.codeium/windsurf/mcp_config.json

command: "npx", args: ["-y", "@appreviewmcp/server"]

Continue[OK]

~/.continue/config.json

"transport": "stdio"

// mcpServers array

Zed[OK]

~/.config/zed/settings.json

"command": "bunx @appreviewmcp/server"

// context_servers key

What devs actually shipped.

commita3f9b1cfeat: add review sentiment to product-health dashboard
“Wired AppReviewMCP into my Claude workflow last week. Now I ask 'what broke in the last update?' and get actual user signal back in seconds. Replaced a 3-hour manual process.”
—@priyabuilds·Senior iOS engineer, Bangalore

commitd72e4a8chore: replace manual review scraping with MCP tool call
“The curl → Claude pipeline is dead simple. Point it at your app ID and you get structured review data in whatever format your LLM needs. No scraping, no rate-limit nightmares.”
—@marcuswrites·Full-stack dev, Berlin

commitc19b5f3docs: add MCP-based review analysis to ADR-042
“Used this for competitive analysis. Queried 15 competitor apps across 6 countries in one Cursor session. The 170-country coverage is the real differentiator — nothing else has it.”
—@elenadev·Product engineer, Warsaw