Jupyter Code Executor MCP Server

A powerful Model Context Protocol (MCP) server that enables AI assistants to execute code in multiple programming languages through Jupyter kernels. This server provides stateful and stateless code execution capabilities with session management, automatic cleanup, and comprehensive data analysis tools.

✨ Key Features

🚀 Multi-Language Code Execution

• Universal Kernel Support: Execute code in any language with an installed Jupyter kernel (Python, R, Julia, etc.)
• Dual Execution Modes:
  • Stateful Sessions: Maintain variable context across multiple executions within the same session
  • Stateless Execution: One-off code execution without persistent context

🎯 Built-in Data Analysis Capabilities

• Pre-configured Environment: Includes essential data science libraries (Pandas, NumPy, Scikit-learn, Seaborn, etc.)
• Expert Prompt Templates: Built-in data analyst prompt with best practices for EDA, ML, and visualization
• Professional Visualizations: Seaborn and Matplotlib integration for publication-quality charts

📋 Prerequisites

• Python: >= 3.11
• uv: Modern Python package installer (recommended)
• Jupyter: Jupyter kernels for your target languages

🔧 Installation

Using uv (Recommended)

# Install the package
uv pip install jupyter-code-executor-mcp-server

# Or install from source
git clone https://github.com/twn39/jupyter-code-executor-mcp-server.git
cd jupyter-code-executor-mcp-server
uv sync
uv run jupyter-code-executor-mcp-server

Using pip

pip install jupyter-code-executor-mcp-server

🚀 Quick Start

Starting the Server

# Start with default settings
uv run jupyter-code-executor-mcp-server

# Start with custom configuration
uv run jupyter-code-executor-mcp-server \
  --port 8080 \
  --data_dir ~/my-data \
  --output_dir ~/my-output \
  --session-timeout 1200

Command Line Options

Option	Short	Description	Default
--port	-p	Server listening port	5010
--data_dir	-d	User data directory	~/data
--output_dir	-o	Code output directory	~/output
--session-timeout	-st	Session timeout (seconds)	600

Testing the Server

Use the MCP Inspector to interact with the server:

pnpx @modelcontextprotocol/inspector

📚 Available MCP Tools

1. list_kernels

List all installed and available Jupyter kernels on your system.

Returns: String containing kernel names and display names

Example Response:

Available Jupyter Kernels:
Format: [Kernel Name] - [Display Name]
----------------------------------------
- python3: Python 3
- ir: R
- julia-1.9: Julia 1.9

2. execute_code

Execute code in a specified Jupyter kernel with optional session persistence.

Parameters:

• code (string): The code to execute
• kernel_name (string): Target kernel name (use list_kernels to find available kernels)
• session_id (string, optional): Session identifier for stateful execution

Returns: Execution result including kernel name, exit code, and output

Stateless Example (Python):

# Execute code without session persistence
code = """
import pandas as pd
df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
print(df.describe())
"""

Stateful Example (Python with session):

# First execution - create variables
session_id = "data-analysis-001"
code1 = """
import pandas as pd
df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
"""

# Second execution - reuse variables from session
code2 = """
# df is still available from previous execution
print(df.describe())
print(f"Shape: {df.shape}")
"""

R Example:

# Execute R code
code = """
data <- data.frame(x = 1:10, y = rnorm(10))
summary(data)
plot(data$x, data$y)
"""
kernel_name = "ir"

📖 Available Prompts

data_analyst_prompt

A comprehensive prompt template for AI-powered data analysis featuring:

• Statistical Analysis: EDA, hypothesis testing, attribution analysis, time series
• Machine Learning: Supervised/unsupervised learning, model evaluation, hyperparameter tuning
• Deep Learning: Neural networks, CNNs, RNNs/LSTMs with TensorFlow/PyTorch
• Professional Visualization: Publication-quality charts with Seaborn/Matplotlib
• Iterative Workflow: Think → Act → Observe → Interpret cycle

This prompt transforms your AI assistant into an expert data scientist with code execution capabilities.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

Development Setup

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Built with FastMCP
• Powered by Jupyter
• Execution engine: AutoGen

---

Note: This server requires properly configured Jupyter kernels for the languages you want to execute. Install kernels using commands like pip install ipykernel (Python), install.packages('IRkernel') (R), or following the respective kernel installation guides.