MCP 服务器

AppDataLayer MCP Server

An app intelligence query engine that enables analysis of over 1 billion reviews from Google Play and the Apple App Store. It provides tools for sentiment analysis, keyword rankings, competitive comparisons, and time-series forecasting across 250,000+ apps.

README

appdatalayer-mcp

Open-source TypeScript SDK for the AppDataLayer MCP Server — the App Intelligence Query Engine powered by the Model Context Protocol.

Analyze 1B+ app reviews across Google Play and the Apple App Store. Semantic topic search, sentiment analysis, keyword rankings, competitive analysis, and time-series forecasting — all through a single typed API.

Installation

npm install appdatalayer-mcp

Quick Start

As a TypeScript/JavaScript SDK

import { AppDataLayerClient } from "appdatalayer-mcp";

const client = new AppDataLayerClient({ apiKey: "sk_live_..." });

// Get app metadata
const app = await client.getAppOverview("com.instagram.android");
console.log(app.data?.title, app.data?.score);

// Analyze review sentiment
const reviews = await client.analyzeReviews({
  app_id: "com.spotify.music",
  dimensions: ["sentiment", "topic"],
  days: 30,
});

// Semantic topic search
const topics = await client.searchTopics("crashes and bugs");

// Compare apps head-to-head
const comparison = await client.compareApps({
  app_ids: ["com.spotify.music", "com.apple.music"],
});

// Forecast review volume
const forecast = await client.forecastMetric({
  series: [100, 120, 115, 130, 128, 145, 142, 160, 155, 170],
  horizon: 7,
});

await client.disconnect();

As an MCP Server for LLM Agents

Use AppDataLayer directly in Claude Desktop, Cursor, Windsurf, or any MCP-compatible AI agent:

import { generateMcpConfig } from "appdatalayer-mcp";

const config = generateMcpConfig("sk_live_...");
console.log(JSON.stringify(config, null, 2));

Paste the output into your agent's config file:

Client	Config File
Claude Desktop	`~/Library/Application Support/Claude/claude_desktop_config.json`
Cursor	`.cursor/mcp.json`
Windsurf	`~/.windsurf/mcp.json`
VS Code	`.vscode/mcp.json`

Generated config:

{
  "mcpServers": {
    "appdatalayer": {
      "url": "https://mcp.appdatalayer.com/mcp",
      "headers": {
        "APPDATALAYER_API_KEY": "sk_live_..."
      }
    }
  }
}

Available Tools

The SDK provides typed methods for all 22 MCP tools:

Data Querying

Method	Description
`getAppOverview(appId)`	App metadata: title, developer, score, installs, rating histogram
`analyzeReviews(input)`	Aggregate reviews by day/week/month/sentiment/country/score/topic
`getKeywordRankings(input)`	Track app ranking for a search keyword over time
`getTopCharts(input?)`	Latest top charts (free, paid, grossing, new)
`getSimilarApps(appId)`	Apps similar/related to a given app
`getSearchSuggestions(input)`	Autocomplete suggestions from app stores
`getGlobalStats()`	Platform totals: apps tracked, reviews count
`getReviewsByTopics(input)`	Actual review text filtered by topic IDs
`compareApps(input)`	Head-to-head comparison of 2-5 apps
`getTopicTrend(input)`	Track topic volume and sentiment changes over time

Topic Intelligence

Method	Description
`resolveTopics(ids)`	Convert topic IDs → human-readable labels
`searchTopics(query)`	Semantic search over 1M+ review topics
`findAppsByTopics(input)`	Find apps at the intersection of two topic sets

Forecasting

Method	Description
`forecastMetric(input)`	Predict future values using Google TimesFM 2.5

Scraping Operations

Method	Description
`getScrapeJobsOverview()`	Overview of all scraping jobs by type
`listScrapeJobs(input?)`	List and filter individual scrape jobs
`getFailedJobs(input?)`	Get failed/dead jobs with failure reasons
`getJobScheduleStatus(input?)`	Get overdue jobs

Webhooks

Method	Description
`listWebhooks()`	List all webhooks for the authenticated user
`createWebhook(input)`	Create a webhook alert rule
`deleteWebhook(id)`	Delete a webhook
`toggleWebhook(id, active)`	Enable/disable a webhook

Raw Access

Method	Description
`call(toolName, args)`	Call any MCP tool by name
`listTools()`	List all available tools
`listResources()`	List all available resources
`readResource(uri)`	Read a resource (e.g. table schemas)

API Reference

`AppDataLayerClient`

const client = new AppDataLayerClient({
  apiKey: string;            // Required — your API key
  endpoint?: string;         // Default: "https://mcp.appdatalayer.com/mcp"
  timeout?: number;          // Default: 30000ms
});

Return Type

All methods return ToolCallResult<T>:

interface ToolCallResult<T> {
  content: { type: string; text: string }[]; // Raw MCP content
  isError?: boolean;                          // True if the tool returned an error
  data: T | null;                             // Parsed JSON data (null if error)
}

`generateMcpConfig`

import { generateMcpConfig } from "appdatalayer-mcp";

const config = generateMcpConfig(
  apiKey: string,
  endpoint?: string  // Default: "https://mcp.appdatalayer.com/mcp"
);