広告掲載
MCP.so
ログイン

MCP Server Research Tool

@atimmeny27

リポジトリ

基本情報

カテゴリ

データと分析

トランスポート

stdio

作者

atimmeny27

MCP Server Research Tool について

Uses an MCP server, API key, and claude to search thousands of sources and compile any topic into neatly formatted notes in markdown.

#mathgpt#math-solver#math-assistant

MCP Server Research Tool の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "MCP-Server-Research-Tool": {
      "command": "python3",
      "args": [
        "-m",
        "ensurepip",
        "--upgrade"
      ]
    }
  }
}

MCP Server Research Tool のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is MCP-Server-Research-Tool?

MCP-Server-Research-Tool is a long-form research assistant powered by LLMs. It gathers information from primary sources, podcasts, academic PDFs, textbooks, YouTube, and more, then outputs structured markdown files with a detailed topic summary and source list.

How to use MCP-Server-Research-Tool?

Clone the repository, run the installation commands, set your OpenRouter API key (exported in .zshrc or .env), then launch with ./start.sh or manually with python MCP_assistant.py "Topic" and MCP_DURATION=long python send_to_claude.py. Enter a specific research topic and wait up to two minutes for the results.

Key features of MCP-Server-Research-Tool

  • Gathers information from primary sources, podcasts, and academic PDFs
  • Outputs structured markdown with summary and source list
  • Powered by LLMs via OpenRouter API (free key required)
  • Configure research prompt by editing send_to_claude.py lines 18-220
  • Supports manual launch or automated start script
  • Designed for macOS, can be added as a Desktop tool via Automator

Use cases of MCP-Server-Research-Tool

  • Conducting deep, realistic research on a specific historical or technological topic
  • Compiling structured notes from multiple media types (audio, video, text)
  • Creating markdown notes for use in editors like Obsidian
  • Learning about a subject by simulating LLM-driven multi-source investigation

よくある質問

1

What API key is required?

You need a free API key from OpenRouter (or optionally from OpenAI/Anthropic). The key is set as environment variable `OPENROUTER_API_KEY` in `~/.zshrc` or a local `.env` file.

2

How long does research take?

The tool waits up to 2 minutes for the LLM to research the topic. You will see a message like “🔐 Using key: sk-or-v1-2 …” when it is actively researching.

3

What is the output?

The output is a markdown (.md) file containing a detailed summary and a list of sources used. It is best opened in a markdown-friendly editor such as Obsidian.

4

Can I edit the research prompt?

Yes. The instructions sent to Claude are in `send_to_claude.py` around lines 18-220 as a long f-string. You can modify that to change the research behavior.

5

Are there any known limitations?

Wikipedia’s API may be overloaded, causing temporary errors. Spelling errors or niche topics may confuse the LLM. For best results, use a specific, detailed research topic.

コメント

「データと分析」の他のコンテンツ

Shipfound

Shipfound

Shipfound turns Claude Code or Codex into your growth engineer: site fixes for Google and AI search, content as pull requests, and proof of every result.

Instagram MCP Server

HasData

Read public Instagram profiles and post feeds by handle in Claude, Cursor, and other MCP clients. Get follower counts, captions, hashtags, mentions, and engagement counts as structured JSON.

23

YouTube MCP Server

HasData

Hosted YouTube MCP server for search, video metadata, channel data and timestamped transcripts. Connect Claude, Cursor or other MCP clients with HasData; no YouTube API key required.

6

Subtext

Subtext by Fullstory

Session replay, built for agents. Subtext is agentic session review: it captures production sessions of your app and connects them to your coding agent — Claude Code, Cursor, Codex, Devin, your own harness — so it can

9

DPF

DPF Admin

DPF (Data Prompt Forge) is an AI-powered platform for data ingestion, transformation and analytics using natural language. Describe your data and the shape you want and the AI infers the schema and generates the pipeline

広告掲載