MCP Server for AI Summarization
A Model Context Protocol (MCP) server that provides intelligent summarization capabilities for various types of content using Google's Gemini 1.5 Pro model. This server can help you generate concise summaries while maintaining key information from different content formats.
The summarization service is powered by 3MinTop, an AI-powered reading tool that helps you understand a chapter's content in just three minutes. 3MinTop transforms complex content into clear summaries, making learning efficient and helping build lasting reading habits.
Clone this repository
Install dependencies:
pnpm install
Build the project:
pnpm run build
Start the server:
pnpm start
pnpm run dev
to start the TypeScript compiler in watch modesrc/index.ts
to customize server behavior or add new toolsTo integrate this server with a desktop app, add the following to your app's server configuration:
{
"mcpServers": {
"content-summarizer": {
"command": "node",
"args": [
"{ABSOLUTE PATH TO FILE HERE}/dist/index.js"
]
}
}
}
Summarizes content from various sources using the following parameters:
content
(string | object): The input content to summarize. Can be:
type
(string): Content type ("text", "url", "pdf", "epub")maxLength
(number, optional): Maximum length of the summary in characters (default: 200)language
(string, optional): Target language for the summary (default: "en")focus
(string, optional): Specific aspect to focus on in the summarystyle
(string, optional): Summary style ("concise", "detailed", "bullet-points")Example usage:
// Summarize a webpage
const result = await server.invoke("summarize", {
content: "https://example.com/article",
type: "url",
maxLength: 300,
style: "bullet-points"
});
// Summarize a PDF document
const result = await server.invoke("summarize", {
content: pdfBase64Content,
type: "pdf",
language: "zh",
style: "detailed"
});
A dynamic resource that demonstrates basic MCP resource functionality:
greeting://{name}
Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License - see the LICENSE file for details.