视频分析 MCP 服务 | Video Analysis MCP Service

中文 | English

中文版本

这是一个基于Model Context Protocol (MCP)的服务，用于视频内容分析和思维导图生成。

更新日志

v1.0.0 (2025-04-17)

• 初始版本发布
• 支持视频分析功能
• 支持思维导图生成

v1.0.2 (2025-04-18)

• 优化生成总结的功能
• bugfix: 修复生成mindmap.json时，参数错误的问题
• refactor: 重构生成思维导图HTML的逻辑
• doc: 新增调用示例 task.md

正在开发

• 支持字幕文件导出
• AI改写
• 展示流 html页面

功能特点

• 视频分析：从视频URL中提取关键词、摘要和关键时间点
• 思维导图生成：基于视频分析结果生成结构化的思维导图
  • 生成JSON格式的思维导图数据（兼容jsMind库）
  • 生成静态PNG图片格式的思维导图
  • 生成交互式HTML格式的思维导图

环境要求

• Node.js >= 14.0.0
• 所需API密钥
  • BILIGPT_API_KEY：用于视频内容分析（支持多种平台）
    • 获取方式：https://bibigpt.co/r/jSeDX0
    • 支持的视频平台：
      • 国内平台：哔哩哔哩、抖音、快手、小红书、西瓜视频、今日头条、可灵、优酷
      • 海外平台：YouTube、TikTok、Instagram、Lemon8
  • DEEPSEEK_API_KEY：用于思维导图JSON生成

克隆项目

git clone https://github.com/tjh19971228/mcp_video_analysis.git
cd mcp_video_analysis

安装

# 安装依赖
npm install
 或者
pnpm install

配置

创建.env文件并设置必要的API密钥：

BILIGPT_API_KEY=your_biligpt_api_key
DEEPSEEK_API_KEY=your_deepseek_api_key

使用方法

启动服务

# 启动MCP服务
node src/index.js

与Claude集成

要在Claude桌面版中使用此MCP服务，编辑claude_desktop_config.json文件：

{
  "mcpServers": {
    "video-analysis": {
      "command": "node",
      "args": ["your_project_path/src/index.js"],
      "env": {
        "BILIGPT_API_KEY": "your_biligpt_api_key",
        "DEEPSEEK_API_KEY": "your_deepseek_api_key"
      }
    }
  }
}

API使用示例

视频分析

// 调用视频分析服务
const result = await videoAnalysis.analyzeVideo({
  url: "https://www.bilibili.com/video/BVxxxxxx"
});

// 返回结果包含：
// - keywords: 关键词列表
// - summary: 视频摘要
// - keyTimepoints: 关键时间点列表

生成思维导图

// 生成思维导图JSON
const json = await mindmap.generateMindmapJson({
  keywords: result.keywords,
  summary: result.summary,
  keyTimepoints: result.keyTimepoints
});

// 生成思维导图图片
const imagePath = await mindmap.generateMindmapImage({
  json: json,
  outputPath: "./output/mindmap.png" // 可选，指定输出路径
});

// 生成思维导图HTML
const htmlPath = await mindmap.generateMindmapHtml({
  json: json,
  outputPath: "./output/mindmap.html", // 可选，指定输出路径
  title: "视频思维导图" // 可选，指定标题
});

开发者指南

服务使用官方的Model Context Protocol SDK实现，主要组件包括：

• src/index.js：服务入口，使用MCP SDK创建和启动服务
• src/services/videoAnalysis.js：视频分析服务实现
• src/services/mindmap.js：思维导图生成服务实现

许可证

MIT

English Version

This is a service based on the Model Context Protocol (MCP) for video content analysis and mind map generation.

Changelog

v1.0.0 (2025-04-17)

• Initial release
• Video analysis support
• Mind map generation

v1.0.1 (2025-04-18)

• Optimized summary generation
• bugfix: Fixed parameter error when generating mindmap.json
• refactor: Restructured the logic for generating mind map HTML
• doc: Added usage examples in task.md

Under Development

• Support for exporting subtitle files
• AI Rewriting
• Demo flow HTML page

Features

• Video Analysis: Extract keywords, summaries, and key timepoints from video URLs
• Mind Map Generation: Create structured mind maps based on video analysis results
  • Generate mind map data in JSON format (compatible with jsMind library)
  • Generate mind maps as static PNG images
  • Generate interactive HTML mind maps

Requirements

• Node.js >= 14.0.0
• Required API Keys
  • BILIGPT_API_KEY: For video content analysis (supports multiple platforms)
    • Get it from: https://bibigpt.co/r/jSeDX0
    • Supported video platforms:
      • Domestic (China): Bilibili, Douyin, Kuaishou, Xiaohongshu, Xigua Video, Toutiao, Keling, Youku
      • International: YouTube, TikTok, Instagram, Lemon8
  • DEEPSEEK_API_KEY: For mind map JSON generation

Clone Project

git clone https://github.com/tjh19971228/mcp_video_analysis.git
cd mcp_video_analysis

Installation

# Install dependencies
npm install
 or
pnpm install

Configuration

Create a .env file and set the necessary API keys:

BILIGPT_API_KEY=your_biligpt_api_key
DEEPSEEK_API_KEY=your_deepseek_api_key

Usage

Start the Service

# Start the MCP service
node src/index.js

Claude Integration

To use this MCP service with Claude desktop, edit the claude_desktop_config.json file:

{
  "mcpServers": {
    "video-analysis": {
      "command": "node",
      "args": ["your_project_path/src/index.js"],
      "env": {
        "BILIGPT_API_KEY": "your_biligpt_api_key",
        "DEEPSEEK_API_KEY": "your_deepseek_api_key"
      }
    }
  }
}

API Examples

Video Analysis

// Call the video analysis service
const result = await videoAnalysis.analyzeVideo({
  url: "https://www.bilibili.com/video/BVxxxxxx"
});

// The result contains:
// - keywords: List of keywords
// - summary: Video summary
// - keyTimepoints: List of key timepoints

Generate Mind Maps

// Generate mind map JSON
const json = await mindmap.generateMindmapJson({
  keywords: result.keywords,
  summary: result.summary,
  keyTimepoints: result.keyTimepoints
});

// Generate mind map image
const imagePath = await mindmap.generateMindmapImage({
  json: json,
  outputPath: "./output/mindmap.png" // Optional, specify output path
});

// Generate mind map HTML
const htmlPath = await mindmap.generateMindmapHtml({
  json: json,
  outputPath: "./output/mindmap.html", // Optional, specify output path
  title: "Video Mind Map" // Optional, specify title
});

Developer Guide

The service is implemented using the official Model Context Protocol SDK, with the main components including:

• src/index.js: Service entry point, creates and starts the service using the MCP SDK
• src/services/videoAnalysis.js: Video analysis service implementation
• src/services/mindmap.js: Mind map generation service implementation

License

MIT