Jupythunder

Name: Jupythunder
Author: kbsooo

@kbsooo

リポジトリ

Jupythunder について

概要はまだありません

#mathgpt #math-solver #math-assistant

基本情報

カテゴリ

その他

トランスポート

stdio

公開者

kbsooo

投稿者

Byungsoo Kang

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": {
    "jupythunder": {
      "command": "/absolute/path/python",
      "args": [
        "/absolute/path/jupythunder_server.py"
      ]
    }
  }
}

ツール

ツールは検出されませんでした

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 Jupythunder?

Jupythunder is a middleware that connects the LLM chat interface (Claude) with Jupyter Notebook. It allows users to interact with Jupyter Notebook using natural language, converting natural language requests into Python code and executing it remotely.

How to use Jupythunder?

Install prerequisites (Python 3.10+, Jupyter Notebook, Claude for Desktop), clone the repository, and install dependencies. Run the Jupyter notebook server and the MCP server, then configure Claude for Desktop with absolute paths to Python and jupythunder_server.py. Use commands like /notebook:name to create notebooks, /open:name to open existing ones, and natural language instructions for data analysis, code execution, and querying.

Key features of Jupythunder

Natural language code conversion by LLM
Remote Jupyter Notebook connection via URL/token
Code execution and result retrieval
Function/variable queries via /question command
Session state and connection management
Notebook file creation, opening, and saving
Natural language data analysis and visualization
Cell content viewing, selective execution, and sequential execution

Use cases of Jupythunder

Perform data analysis by describing steps in natural language
Automate Python coding without manual notebook interaction
Query and debug notebook functions and variables conversationally
Create and manage .ipynb files remotely through chat
Execute cells selectively or sequentially for iterative workflows

FAQ from Jupythunder

What is Jupythunder and how does it differ from other Jupyter tools?

Jupythunder is a middleware that links Claude (an LLM chat interface) directly to a Jupyter Notebook server, enabling natural language commands for coding and analysis instead of manual notebook editing.

What are the runtime dependencies for Jupythunder?

Python 3.10+, Jupyter Notebook, Claude for Desktop (or another MCP client), and Python packages: mcp[cli]>=1.5.0, requests>=2.31.0, websocket-client>=1.6.0, python-dotenv>=1.0.0, urllib3>=2.0.0.

Where does data live when using Jupythunder?

Data and notebooks reside on the remote Jupyter Notebook server; Jupythunder connects to that server via a provided URL and token, executes code, and saves results there.

What are known limitations or common errors?

The most common error is spawn pythonENOENT when the Python executable path is not found; use absolute paths or python3. Jupyter connection fails if the server is not running or the URL/token is incorrect. Notebook creation may fail due to permissions or invalid names.

What transport and authentication does Jupythunder use?

Jupythunder uses the MCP (Model Context Protocol) for communication with Claude and connects to Jupyter via HTTP/WebSocket using a user-provided URL and token for authentication.