Skip to content

🧪 Tutorial: MIMIC + Jupyter Pipeline

🧪 Tutorial: MIMIC + Jupyter Pipeline

This tutorial shows how to stack two MCPs:

  • MIMIC (conversational AI on the MIMIC-IV clinical database)
  • Jupyter (reproducible notebook analysis)

You’ll learn how to:

  1. Build a pipeline with both tools
  2. Ask natural language questions over MIMIC-IV
  3. Export results into a Jupyter Notebook for reproducible Python analysis

🎥 Video Walkthrough

Prerequisites

# If you prefer other Python package managers, feel free to adapt `pip install X`.

uv init --python 3.10
uv add mcpstack
uv add mcpstack-jupyter
uv add mcpstack-mimic

# To see if the tools are all connected
uv run mcpstack list-tools

🔧 Step 1 — Build with Pipeline W/ MIMIC Default

MIMIC Default is basically using the default MIMIC-IV demo database.

uv run mcpstack pipeline mimic --new-pipeline my_pipeline.json

🔧 Step 2 — Create a Jupyter ToolConfig

Basically, Jupyter MCP works with some sort of connections between the LLM and the Jupyter instance. This is via a URL and a TOKEN. Hence, the need for a ToolConfig.

uv run mcpstack tools jupyter configure \
    --token YOUR_JUPYTER_TOKEN

# This create a `jupyter_config.json` file

🔧 Step 3 — Add To The Tool To The Pipeline

uv run mcpstack pipeline jupyter --to-pipeline my_pipeline.json --tool-config jupyter_config.json

🔧 Step 4 — Compose & Run the Pipeline On Claude Desktop

uv run mcpstack build --pipeline my_pipeline.json --config-type claude

Now you can ask the LLM to operate the MIMIC tool and export results into Jupyter.

📣 Prompt Used During The Demo Video

Hey there! May you extract 50 patients from MIMIC-IV and do a quick reproducible analysis of the data with pandas
for me please?

If you need to add packages, `!uv add <package_name>`.

Tip

Try chaining additional tools (e.g., scikit-longitudinal) to build research-ready clinical workflows.