Run Jupyter notebooks online, without installing anything
No Anaconda downloads, no environment conflicts, no 'works on my machine'. NotebookHub gives you a full JupyterLab — the real thing, not a stripped-down toy — running in your browser on dedicated cloud resources, with storage that persists between sessions.
How it works
From zero to a running notebook in three steps:
- 1 · Create your account and choose an available paid plan
- 2 · Pick a kernel: Python (SciPy stack), R, geospatial (GeoPandas, rasterio, GDAL) or PyTorch
- 3 · Your JupyterLab opens in the browser — upload data or clone a repo, and work
A real JupyterLab, with real resources
Browser-only Python playgrounds are fine for snippets, but they choke on real data work: no filesystem, tiny memory, no long-running cells. Every NotebookHub workspace is a dedicated container with reserved CPU and RAM, a persistent disk, and the complete JupyterLab interface — terminals, extensions and all.
- Persistent disk (10 GB on the trial, up to 100 GB paid) — your work is there tomorrow
- Dedicated resources even on the trial (1 vCPU / 2 GB RAM; paid plans up to 4 vCPU / 32 GB)
- Full terminal access: pip, conda, git, wget
- Prefer an IDE? The same workspace also runs VS Code in the browser
For students, researchers and data teams
Use it to follow a course without fighting local installs, to run analyses from a Chromebook or a locked-down office laptop, or to keep a reproducible environment that your whole team can mirror. When the trial ends, paid plans add more workspaces, much more memory and team features.
Frequently asked questions
Is it really free to try?
New trials are temporarily disabled. Paid plans start at US$16/month (charged as $15,000 CLP), subject to stock.
Can I install my own packages?
Yes — pip and conda work normally, and installs persist because the disk is persistent.
Can I open my existing .ipynb files?
Yes, upload them directly in JupyterLab or clone your git repository from the built-in terminal.
What Python version and libraries are preinstalled?
Workspaces start from the Jupyter Docker Stacks images (SciPy stack: numpy, pandas, matplotlib, scikit-learn…), with R, geospatial and PyTorch variants available.
Your notebook is one signup away
JupyterLab in your browser in under a minute. Paid plans subject to availability.