langchain-widget 0.0.1

A widget to run in Jupyter Notebook.

LangChain Widget

Run a LangChain agent in Jupyter notebooks.

An anywidget + React chat UI for LangChain tool-calling chat models, built for notebooks and designed to compose with other Jupyter widgets.

• Works in JupyterLab / Notebook (and VS Code notebooks)
• Tool calling out of the box (your tools, your model)
• Easy to embed alongside plots, tables, and domain widgets

Installation

pip install langchain-widget

Quick start

In a notebook, create a LangChainWidget with your chat model and tools:

from langchain_widget import LangChainWidget
from langchain_openai import ChatOpenAI
from langchain_core.tools import tool

chat_model = ChatOpenAI(model="gpt-4o-mini")

@tool
def add(a: int, b: int) -> int:
    "Add two integers."
    return a + b

w = LangChainWidget(
    chat_model=chat_model,
    tools=[add],
    system_prompt="You are a helpful scientific assistant.",
)
w

Here is how it looks in a Jupyter notebook:

Documentation

Full documentation at: https://langchain-widget.readthedocs.io/

Development

See langchain-widget/docs/source/development.rst.

Contact

• Xing Wang [e-mail]

License

MIT