• Building a Chat Interface for Jupyter with LangChain
  • What you’ll learn
  • Project files
  • Docker: build + run JupyterLab
  • Running the notebook
  • Provider configuration

Building a Chat Interface for Jupyter with LangChain¶

This folder demonstrates how to build a chat interface for Jupyter notebooks using LangChain and ipywidgets, with an optional JupyterLab extension for production use.

What you’ll learn¶

• Building a basic chat UI with ipywidgets
• Integrating LangChain for AI-powered responses
• Managing conversation state and memory
• Making chat context-aware of notebook variables
• Handling async operations in Jupyter
• Building a production JupyterLab extension

Project files¶

• jupyter_chat_extension.ipynb: complete tutorial notebook with runnable code
• blog_jupyter_chat_langchain.md: comprehensive guide covering both widget-based and extension approaches
• jupyterlab_context_chat/: full JupyterLab extension (sidebar chat with notebook context)
• requirements.txt: Python dependencies
• Dockerfile: installs deps for JupyterLab
• docker-compose.yml: starts JupyterLab with the repo mounted at /app
• .env.example: template for environment variables (copy to .env)

Two approaches¶

This tutorial shows two ways to add chat to Jupyter:

1. In-Notebook Chat (ipywidgets)¶

Quick and easy - add chat directly in your notebook using ipywidgets. See jupyter_chat_extension.ipynb.

Pros:

• Simple to implement
• No extension installation needed
• Works in any Jupyter environment

Cons:

• Chat lives inside notebook cells
• Limited UI customization
• Restarts with kernel

2. JupyterLab Extension (recommended for production)¶

A proper JupyterLab sidebar extension with backend endpoint. See jupyterlab_context_chat/.

Pros:

• Professional sidebar UI
• Persistent across notebooks
• Server-side LLM calls (API keys stay secure)
• Captures notebook context automatically
• Distributable as Python package

Cons:

• Requires extension installation
• More complex setup

For detailed comparison and full tutorial, see blog_jupyter_chat_langchain.md.

Extension setup: jupyterlab_context_chat/SETUP.md

Docker: build + run JupyterLab¶

From this directory:

cd tutorials/Jupyter_Extension_Langchain
cp .env.example .env
# Edit `.env` and set your API key.
docker compose up --build

Then open:

• http://localhost:8888/lab

Notes:

• The compose file disables the Jupyter token/password for convenience. Don’t use this on an untrusted network.
• The repo is mounted into the container (/app), so edits on your host are reflected instantly.

Running the notebook¶

Open jupyter_chat_extension.ipynb and run it top-to-bottom.

Notes:

• Running this notebook will call your configured LLM provider and may incur costs
• The notebook demonstrates progressively more advanced chat features
• Context-aware examples require creating variables in the notebook first

Provider configuration¶

Set env vars in .env (loaded via docker-compose.yml) and restart the container. See .env.example for the supported variables.

Changelog¶

• 2026-03-01: Initial release