Search Fetch

Provides a notes-based MCP server with note resources, create/fetch tools, and note summarization prompts.

typescript

0

GitHub Stars

typescript

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nexon33-search-fetch-server-mcp": {
      "command": "node",
      "args": [
        "/path/to/search-fetch-server/build/index.js"
      ]
    }
  }
}

This MCP server provides a simple notes system implemented with Model Context Protocol. It exposes resources for text notes, tools to create and fetch notes and URLs, and prompts to generate note summaries, enabling you to build interactive, AI-assisted note workflows.

How to use

You interact with the search-fetch server through an MCP client by calling its resources and tools. Create notes, fetch URL content, perform searches, and request summaries of your stored notes. Use note URIs to access individual notes and take advantage of the built-in prompts to generate concise overviews of your notes for consumption by an AI model.

Create a new note with the create_note tool by providing a title and content.
Fetch content from a URL using fetch_url; you can opt to convert the content to Markdown with Puppeteer or receive raw content via Axios.
Perform a DuckDuckGo search with duckduckgo_search by supplying a query; results are returned as JSON.
Ask summarize_notes to generate a structured summary of all stored notes, with embedded note content included as resources.

How to install

Prerequisites: you need Node.js installed on your machine to run the MCP server locally.

npm install
npm run build
npm run watch

Additional sections

Development notes and how to run locally are covered in the development and installation steps. The server is designed to run as a standard MCP stdio process, starting via a Node.js command that points to the built index. Debugging can be aided by the MCP Inspector tool, which provides a browser-based interface for debugging MCP endpoints.

Available tools

create_note

Create new text notes by supplying a title and content; stores the note in server state.

fetch_url

Fetch content from a URL and optionally convert to Markdown using Puppeteer; returns the content as Markdown when Puppeteer is used, otherwise returns the raw content.

duckduckgo_search

Perform a DuckDuckGo search and return the results as a JSON string.

summarize_notes

Generate a structured summary of all stored notes, including embedded note contents as resources, suitable for feeding into an LLM.