Sensor Tower MCP Server Pro
Provides 23 production-tested tools for mobile app intelligence, including app metadata, downloads, revenue, rankings, and search, with multi-token failover for uninterrupted API access.
README
Sensor Tower MCP Server Pro
A pure Node.js implementation of the Model Context Protocol server for Sensor Tower APIs. Access mobile app intelligence data directly in MCP-compatible tools like Cursor, Claude Desktop, and Cline.
Features
- 23 API Tools - Production-tested and verified working tools only
- App Intelligence - Metadata, rankings, downloads, revenue estimates
- Search & Discovery - Find apps and publishers by name or description
- Market Analysis - Category rankings, featured apps, competitor insights
- Store Marketing - Keywords, reviews, featured apps data
- Multi-Token Failover - Automatic switch to backup tokens when quota exhausted
- No Python Required - Pure Node.js implementation
Quick Start
# Run directly with npx (recommended)
npx @feedmob/sensor-tower-mcp-pro
# Or install globally
npm install -g @feedmob/sensor-tower-mcp-pro
sensor-tower-mcp-pro
Requirements
- Node.js 18+
- Sensor Tower API token
MCP Configuration
Cursor / Claude Desktop / Cline
Add to your MCP settings:
{
"mcpServers": {
"sensortower": {
"command": "npx",
"args": ["-y", "@feedmob/sensor-tower-mcp-pro"],
"env": {
"SENSOR_TOWER_API_TOKEN": "primary_token",
"SENSOR_TOWER_API_TOKEN_BACKUP": "backup_token"
}
}
}
}
Multi-Token Failover
When a token's API quota is exhausted (429/403 error), the server automatically switches to the next available backup token. This ensures uninterrupted API access.
Environment Variables
| Variable | Description |
|---|---|
SENSOR_TOWER_API_TOKEN |
Primary API token (required) |
SENSOR_TOWER_API_TOKEN_BACKUP |
First backup token |
SENSOR_TOWER_API_TOKEN_2 |
Second backup token |
SENSOR_TOWER_API_TOKEN_3 |
Third backup token (up to 10) |
Example with Multiple Tokens
SENSOR_TOWER_API_TOKEN=token1 \
SENSOR_TOWER_API_TOKEN_BACKUP=token2 \
SENSOR_TOWER_API_TOKEN_2=token3 \
npx @feedmob/sensor-tower-mcp-pro
API Token
Get your API token from Sensor Tower Account Settings.
Available Tools (23 Total)
App Analysis (13 tools)
| Tool | Description |
|---|---|
get_app_metadata |
App details, ratings, categories |
get_download_estimates |
Download trends and estimates |
get_revenue_estimates |
Revenue data and forecasts |
get_usage_active_users |
Active users data |
top_in_app_purchases |
Top in-app purchases |
version_history |
App version history |
get_category_history |
Category ranking history |
compact_sales_report_estimates |
Compact download/revenue estimates |
category_ranking_summary |
Today's category ranking summary |
app_analysis_retention |
Retention analysis data ⭐ |
downloads_by_sources |
Downloads by source (organic/paid) |
app_analysis_demographics |
Demographic analysis |
app_update_timeline |
App update history |
⭐ Fixed:
app_analysis_retentionnow only supportsall_timeorquarterlyfor date_granularity.
Market Analysis (1 tool)
| Tool | Description |
|---|---|
get_store_summary |
App store summary statistics |
Store Marketing (2 tools)
| Tool | Description |
|---|---|
get_featured_today_stories |
Featured Today stories |
get_featured_apps |
Featured apps on App Store |
Search & Discovery (3 tools)
| Tool | Description |
|---|---|
get_publisher_apps |
Apps by publisher |
get_unified_publisher_apps |
Unified publisher apps |
get_app_ids_by_category |
App IDs by category |
Utilities (4 tools)
| Tool | Description |
|---|---|
get_country_codes |
Available country codes |
get_category_ids |
Platform category IDs |
get_chart_types |
Ranking chart identifiers |
health_check |
Service health status |
Command Line Options
npx @feedmob/sensor-tower-mcp-pro [OPTIONS]
Options:
--transport <type> Transport type: stdio (default) or http
--port <port> HTTP port (default: 8666, only for http transport)
--token <token> API token (or set SENSOR_TOWER_API_TOKEN env var)
--backup-token <tok> Backup API token for failover
--help, -h Show help message
Examples
# Run with stdio transport (for MCP clients)
SENSOR_TOWER_API_TOKEN=your_token npx @feedmob/sensor-tower-mcp-pro
# Run with HTTP transport
npx @feedmob/sensor-tower-mcp-pro --transport http --port 8666
Development
# Install dependencies
npm install
# Build
npm run build
# Run locally
npm start
Links
License
MIT License
Version History
v1.4.0 (2026-02-03)
Production-Ready Release - Tested and Verified Tools Only
This release removes all failing tools based on comprehensive testing, keeping only the 23 tools that are verified to work in production.
Changes:
- ✅ Removed 9 failing tools based on actual API testing:
get_creatives,get_impressions,impressions_rank(App Analysis - HTTP 422 or empty data)get_featured_creatives(Store Marketing - Server bug)games_breakdown,top_apps,top_apps_search(Market Analysis - Unofficial API failures)sales_reports,unified_sales_reports(Your Metrics - Account connection required or server bugs)
- ⭐ Fixed
app_analysis_retentionparameters:- Changed
date_granularityto only supportall_timeorquarterly(was incorrectly allowing daily/weekly/monthly) - Added proper validation for
end_daterequirement when using quarterly mode - Updated documentation with usage examples
- Changed
- ✅ Quality improvements:
- All 23 remaining tools are production-tested and verified
- Removed 200+ lines of non-working code
- Updated health_check to report correct tool count (23)
- Comprehensive documentation updates
Tool Count: 32 → 23 tools (production-tested only)
What's Working:
- ✅ All app metadata and information tools
- ✅ Download and revenue estimation tools
- ✅ Ranking and category analysis tools
- ✅ User behavior analysis tools (with corrected parameters)
- ✅ Publisher and discovery tools
- ✅ Featured content tools
- ✅ All utility tools
Migration Guide:
- If using
app_analysis_retention: Updatedate_granularitytoall_timeorquarterlyonly - If using removed tools: These tools failed in testing and should not be used
- All remaining 23 tools are stable and production-ready
Testing: Based on comprehensive API testing documented in "SensorTower MCP 工具更新文档 2026-02-03"
v1.3.0 (2026-02-03)
Major Optimization Release - API Alignment
This release aligns the tool collection with official Sensor Tower API documentation, removing deprecated endpoints and improving reliability.
Changes:
- ✅ Removed 11 deprecated tools that return 404 errors:
search_entities,get_category_rankings(Search)get_top_and_trending,get_top_publishers,usage_top_apps,top_creatives(Market Analysis)get_keywords,get_reviews,research_keyword(Store Marketing)analytics_metrics,sources_metrics(Your Metrics)
- ✅ Fixed parameter issues:
- Made
unified_app_idsrequired inunified_sales_reports - Clarified
app_idrequirement inget_featured_creatives
- Made
- ✅ Added documentation warnings:
- Permission requirements for
sales_reports - Server error warnings for
get_store_summary - Unofficial API warnings for shadow endpoints (
top_apps,top_apps_search,games_breakdown)
- Permission requirements for
- ✅ Code quality improvements:
- 343 lines of deprecated code removed
- TypeScript compilation optimized
- Better error messages and documentation
Tool Count: 43 → 23 tools (production-tested only)
Migration Guide:
- If you were using removed tools, check the official Sensor Tower API docs for alternative endpoints
- Shadow API tools (marked with ⚠️) still work but may change without notice
v1.2.13 (2026-01-30)
- Fixed API endpoint paths and date parameter handling
- Improved error handling
v1.2.12 (2026-01-30)
- Initial npm release
- 43 API tools with multi-token failover support
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。