Announcing: Jupyter Notebooks integration.

You can now embed Foxglove directly into Jupyter Notebooks, JupyterLab, Google Colab, and even VS Code. When you’re analyzing logged data, the Foxglove SDK offers a widget that gives you full access to Foxglove’s visualization without ever leaving your notebook.

No more launching a separate app or opening a separate tab. Just import the package, log your messages, and inspect your data using your preferred layout.

Why we built this.

Experimentation often lives inside a notebook, especially in robotics and ML workflows. But until now, visualizing real-world robotics data inside one has been a clunky process. You either saved to disk and opened Foxglove separately, or you reached for simpler tools that don’t support robotics-native formats and schemas.

We wanted to remove that friction. Using the Foxglove SDK, you can now pass data directly from your notebook kernel to a Foxglove viewer running inside the notebook UI. It works with the same layouts you already use in the full app, supporting buffered data. Under the hood, it’s powered by Anywidget, making it portable across notebook environments.

Visualizing buffered data.

Import and use the FoxgloveViewer widget alongside the data streaming APIs from the Foxglove SDK. Log messages in memory and send them in one go.

Example

import math
import json

import foxglove
from foxglove.channels import SceneUpdateChannel
from foxglove.schemas import (
    Color,
    CubePrimitive,
    SceneEntity,
    SceneUpdate,
    Vector3,
)

# load any layout from foxglove
with open('example_layout.json', 'r') as file:
    layout = json.load(file)

scene_channel = SceneUpdateChannel("/scene")

# create a foxglove viewer to collect logged messages
viewer = foxglove.create_notebook_viewer(orgSlug="foxglovehq", layout=layout)

for t in range(10*30):
    timestamp = t * 0.033
    size = abs(math.sin(timestamp)) + 1

    scene_channel.log(
        SceneUpdate(
            entities=[
                SceneEntity(
                    cubes=[
                        CubePrimitive(
                            size=Vector3(x=size, y=size, z=size),
                            color=Color(r=1.0, g=1.0, b=0, a=1.0),
                        )
                    ],
                ),
            ]
        ),
        log_time=int(timestamp*1e9)
    )

# displayer logged data using Foxglove
viewer.show()

How it works.

We built this on top of Anywidget, which abstracts over the complexity of syncing Python and frontend state across environments. By leveraging our Python SDK, this gives you a live Foxglove instance running inside the notebook which is capable of rendering dynamic 3D scenes, timeseries plots, images, and any custom schema you define.

There’s no MCAP export step, no local file management, and no server process to spin up.

Foxglove is the most comprehensive and performant visualization tool for Physical AI. With this new integration, we’re extending observability even closer to where robotics developers are experimenting.

Sign up and get started today.