---
id: rerun-notebook
version: "0.36.0"
license: MIT OR Apache-2.0
license_treatment: permissive
maintenance: active
---

# rerun-notebook — Implementation helper for running rerun-sdk in notebooks

License: permissive · Maintenance: active · Downloads: 121.7K/mo

## What it is and what it does

rerun-notebook is a support package that integrates visualization capabilities into Jupyter notebooks through anywidget. It bundles a WebAssembly-compiled viewer and JavaScript glue code, allowing interactive visualizations of structured data within notebook cells. The package is typically installed indirectly as a dependency rather than used directly.

The package handles asset delivery in multiple modes: by default it loads assets from app.rerun.io, but can be configured via the RERUN_NOTEBOOK_ASSET environment variable to inline assets, serve them locally from a background thread, or load from a custom URL. It depends on anywidget, ipykernel, and jupyter-ui-poll to integrate with the notebook environment.

Use it for:

- Visualize structured data logged during exploratory analysis in a Jupyter notebook.
- Build interactive debugging workflows where you log intermediate results and inspect them inline.
- Create shareable notebook reports that include rich, interactive visualizations.
- Prototype data pipelines with live feedback on logged events.
- Serve notebook-based visualizations in restricted network environments by configuring alternative asset modes.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Provides Jupyter notebook integration for interactive visualization of logged data within notebook cells through an anywidget-based viewer.

Yes, if you need notebook visualization support. Install it via the recommended extra rather than directly. It is actively maintained with no known vulnerabilities and adds minimal install friction. The permissive dual license (MIT or Apache-2.0) poses no restrictions. The main consideration is network access: by default it loads assets from app.rerun.io, but alternative asset modes can be configured.

## Install

pip install rerun-notebook
uv add rerun-notebook
poetry add rerun-notebook

## Installing rerun-notebook

Before you install: Low friction installation as a pure Python wheel. Actively maintained with a recent release 4 days ago. Designed as a support package typically installed via an extra rather than directly.

License in practice: Dual-licensed under MIT or Apache-2.0, both permissive licenses allowing commercial and private use with minimal restrictions.

Quickstart:

# Install as a dependency (typically via extra)
pip install rerun-notebook

# In a Jupyter notebook:
import rerun_notebook
# Typically used indirectly through anywidget, ipykernel, jupyter-ui-poll integration

Requires Jupyter/IPython environment; asset loading defaults to https://app.rerun.io unless RERUN_NOTEBOOK_ASSET environment variable is set before import.

Verify before relying:

- Whether the memory leak mentioned for inlined assets has been resolved in version 0.36.0
- Performance characteristics of serve-local asset mode in different notebook environments
- Compatibility with specific Jupyter versions or notebook server configurations
- Whether direct installation is ever necessary versus installation through the recommended extra

## Package facts

- License: MIT OR Apache-2.0 (permissive)
- Python support: unspecified
- Install friction: low
- Maintenance: active
- Downloads: 121.7K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

jupyter notebook visualization widget, notebook integration visualization, interactive data viewer jupyter, notebook logging visualization, jupyter anywidget viewer, jupyter-widget, visualization

[View on SkillFed](https://skillfed.io/packages/rerun-notebook) · [View on PyPI](https://pypi.org/project/rerun-notebook/)