MCP 服务器

X-Trend Intelligence MCP Server

Provides social intelligence tools for AI agents to analyze competitor sentiment, trends, and brand mentions from X/Twitter data.

README

X-Trend Intelligence MCP Server

A production-ready MCP (Model Context Protocol) server that provides deterministic social intelligence tools for AI agents (Claude, Cursor, ChatGPT). Instead of a dashboard, it exposes tools that AI agents can call to get competitor sentiment, trend analysis, and brand monitoring from X/Twitter data.

Features

8 MCP Tools for social intelligence:
1. analyze_competitor_sentiment — Score sentiment of posts mentioning a competitor
2. track_brand_mentions — Track brand mention volume and engagement over time
3. detect_trending_topics — Get X trends and identify rising topics
4. compare_share_of_voice — Compare mention counts and engagement for 2-5 competitors
5. identify_complaints — Find categorized complaints about a brand
6. monitor_keyword_frequency — Track keyword appearance frequency over time
7. analyze_influencer_reach — Analyze an X user profile and their engagement
8. export_intelligence_report — Generate a combined markdown intelligence report
Deterministic sentiment scoring — No LLM dependency. Uses a 50+ word positive/negative keyword lexicon.
MCP-native — AI agents call tools directly, no dashboard needed.
Vercel-ready — Deploys as serverless functions with Streamable HTTP transport.
BYO X API token — Users bring their own X API v2 bearer token.

Quick Start

Prerequisites

Node.js 20+
An X API v2 bearer token (get one at developer.x.com)
X API access tier with recent search endpoint (Basic or higher)

Installation

git clone <repo-url>
cd x-trend-intelligence-mcp
npm install
npm run build

Local Development

npm run build
# Start the server (requires a Node.js HTTP server wrapper for local testing)
node dist/index.js

Deploy to Vercel

vercel deploy

The MCP endpoint will be at https://your-project.vercel.app/api/mcp and the health check at /api/health.

Usage

Connecting with Claude Desktop

Add to your Claude Desktop config:

{
  "mcpServers": {
    "x-trend-intelligence": {
      "url": "https://your-project.vercel.app/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_X_API_BEARER_TOKEN"
      }
    }
  }
}

Connecting with Cursor

Add to Cursor's MCP settings:

{
  "mcpServers": {
    "x-trend-intelligence": {
      "url": "https://your-project.vercel.app/api/mcp",
      "headers": {
        "x-api-key": "YOUR_X_API_BEARER_TOKEN"
      }
    }
  }
}

Authentication

Provide your X API bearer token via one of:

Authorization: Bearer <token> header
x-api-key: <token> header
x-api-bearer: <token> header

Health Check

curl https://your-project.vercel.app/api/health

Tool Reference

analyze_competitor_sentiment

{
  "tool": "analyze_competitor_sentiment",
  "arguments": {
    "competitor_name": "competitor",
    "days_back": 7,
    "max_results": 100
  }
}

Returns sentiment score (-1 to +1), positive/negative/neutral percentages, engagement stats, and top posts.

track_brand_mentions

{
  "tool": "track_brand_mentions",
  "arguments": {
    "brand_name": "YourBrand",
    "time_window": "7d",
    "max_results": 100
  }
}

Returns daily mention counts, trend direction, engagement stats, and top posts.

detect_trending_topics

{
  "tool": "detect_trending_topics",
  "arguments": {
    "location_id": "1",
    "top_n": 10,
    "verify_volume": true
  }
}

Returns trending topics with verified post volume and trend change percentages.

compare_share_of_voice

{
  "tool": "compare_share_of_voice",
  "arguments": {
    "competitors": ["Brand1", "Brand2", "Brand3"],
    "days_back": 7
  }
}

Returns share of voice percentages, mention counts, and sentiment for each competitor.

identify_complaints

{
  "tool": "identify_complaints",
  "arguments": {
    "brand_name": "YourBrand",
    "days_back": 7,
    "max_results": 200
  }
}

Returns categorized complaints (Performance, Bugs, Customer Service, Pricing, etc.) with top posts.

monitor_keyword_frequency

{
  "tool": "monitor_keyword_frequency",
  "arguments": {
    "keyword": "AI agents",
    "days_back": 7,
    "max_results": 500
  }
}

Returns daily counts, peak day, related hashtags, and sample posts.

analyze_influencer_reach

{
  "tool": "analyze_influencer_reach",
  "arguments": {
    "username": "elonmusk",
    "max_tweets": 100
  }
}

Returns profile info, engagement rates, top performing content, and estimated reach.

export_intelligence_report

{
  "tool": "export_intelligence_report",
  "arguments": {
    "brand_name": "YourBrand",
    "competitors": ["Competitor1", "Competitor2"],
    "days_back": 7,
    "include_complaints": true,
    "include_share_of_voice": true
  }
}

Returns a structured markdown report combining brand mentions, competitor sentiment, share of voice, and complaints.

Sentiment Scoring

Sentiment is scored using a deterministic keyword lexicon approach:

Positive lexicon (50 terms): love, great, amazing, awesome, excellent, perfect, recommend, best, fantastic, etc.
Negative lexicon (50 terms): hate, terrible, awful, worst, broken, sucks, disappointed, scam, useless, buggy, etc.
Formula: score = (positive_count - negative_count) / total_posts
Range: -1.0 (fully negative) to +1.0 (fully positive)
Labels: positive (>0.1), negative (<-0.1), neutral (-0.1 to 0.1)

X API Endpoints Used

Endpoint	Usage
`GET /2/tweets/search/recent`	Search recent posts (last 7 days)
`GET /2/users/by/username/:username`	User profile lookup
`GET /2/users/:id/tweets`	User timeline
`GET /2/tweets/:id`	Single tweet
`GET /2/trends`	Trending topics (requires elevated access)