A spatial Jupyter client that extends notebooks into visionOS
Pulto connects to user-provided Jupyter servers and extends Jupyter notebooks with spatial outputs — USDZ models, point clouds, and Gaussian splats — rendered natively on Apple Vision Pro.
Built for the Current Pulto App
Pulto focuses on notebook viewing, Jupyter connection, and inline rendering of the outputs data teams already use — extended with USDZ, point cloud, and Gaussian splat support.
Notebooks.
Native to visionOS.
Markdown, code cells, and saved outputs in one spatial window.
Your Jupyter server. Your rules.
output formats rendered inline
3D assets, in your space
Point clouds.
With real depth.
Spatial previews built for Apple Vision Pro.
sign-ins to explore the bundled samples
local notebooks stay on device by default
Every output. Rendered inline.
Keychain-secured.
Jupyter credentials never leave the system keychain.
Notebook Workflow
Pulto is designed around the path from notebook selection to readable, spatially aware output review.
Open a Sample
Use the bundled notebook viewer guide to see Pulto without setting up any server.
Connect Jupyter
Enter a server URL and token or password for a Jupyter server you control.
Choose a Notebook
Browse available notebooks and load the one you want to inspect.
Review Outputs
Read markdown, tables, charts, images, point clouds, Gaussian splats, and USDZ previews inline.
For Apple Vision Pro
Pulto is a visionOS app in the Developer Tools and Productivity space, built for notebook review and spatial data inspection.
No Required Login for Samples
Bundled sample notebooks can be viewed without signing into an external service.
Optional Network Connections
Jupyter access is configured by the user and may use HTTPS or local-network discovery.
Next: Notebooks as Live USD Scenes
Spatial Notebook USD, in development — cell results become a live, provenance-tracked USD scene graph you review, refine, and eventually touch.
USD-Backed Outputs
Cell outputs become USD scene objects at stable paths, rendered with RealityKit and exportable as USD or USDZ.
Reactive Session Graph
Cells, executions, outputs, and variables form one synchronized graph — the scene, notebook card, and session ornament all react to the same state.
Provenance & History
Every object in space links back to its notebook, cell, and execution — reruns append history instead of erasing it.
Bidirectional Spatial Computing
The long-term goal: interact with objects in space, the kernel recomputes, and the scene and notebook update together.
Be in the loop for launch
Want early access, launch news, or to talk about using Pulto with your team's spatial data? Send us a note and we'll keep you posted.
Get Launch UpdatesWe only use your email for Pulto updates — no lists shared, ever.