Your next analysis.Without starting over.
An environment ready for Python, R and geospatial data. Work in your browser, pause, and return to the files you left behind.
Paid plans subject to stock. New trials are temporarily disabled.
import pandas as pd df = pd.read_parquet("sales.parquet")df.groupby("region").revenue.sum().plot.bar()Python · pandas · matplotlib
Interactive example. Does not execute code or show live activity.
Pick up where you left off
Pausing keeps your files. No need to start over.
Know what you are paying for
Visible CPU, RAM and disk. Stock checked before payment.
Real people behind the platform
Operated by Aquedra SpA. Support in English and Spanish.
From an idea to your environment
From zero to running, in three steps
An account, a plan with available stock, and the environment your project needs.
Create your account
Register with email and password. Then choose an available plan from Account.
Pick your environment
JupyterLab or VS Code, with Python (SciPy stack), R, geospatial (GeoPandas, rasterio, GDAL) or PyTorch preinstalled.
Start working
Your workspace opens in the browser. Upload data or clone a repo — and it is all still there tomorrow.
Pause when you finish and resume later with your files. Pausing does not cancel a monthly subscription; wallet accounts stop consuming credit while paused.
You are in control
One panel, your whole lab
Create, pause, resume and delete workspaces. We run the cluster; you keep your focus on the data.
Example dashboard. Your account shows your actual resources and status.
Built for two kinds of work
We're not trying to be every cloud notebook. These are the two jobs we do better than a generic platform.
Geospatial and data teams
Consultancies and public agencies that work with rasters, vectors and time series, and lose days rebuilding the same environment on every machine.
- A geospatial image with the stack already installed — no GDAL compiling
- Enough RAM and disk for real datasets, on a persistent volume
- Publish your layers with GISDeploy, our OGC server, from the same account
Courses and classrooms
Teach without spending the first two sessions fixing laptop installs. Every student gets an identical environment in the browser.
- Same environment for everyone, from any laptop or Chromebook
- No cluster to operate and no JupyterHub to maintain
- Group plans for a semester — write to us and we set it up
How it compares
Where NotebookHub helps — and where it doesn't.
| What matters | NotebookHub | Free cloud notebooks | Local install |
|---|---|---|---|
| Your files after closing | Kept on a persistent disk | Wiped with the session | Kept — on one machine only |
| Environment setup | Preinstalled, ready to run | pip install on every session | Hours, and it drifts |
| CPU and RAM | Reserved for you (Guaranteed QoS) | Shared and variable | Whatever your laptop has |
| VS Code in the browser | Included, same workspace | Not available | Local only |
| Private data | Isolated container per user | Third-party terms apply | Private, but hard to share |
Honest caveat: for free GPU training, Colab still wins — we don't sell GPU workspaces yet. Read the full comparison →
Clear resources. Your choice.
Plans
Choose by resources. Availability is confirmed before payment.
New trials are temporarily disabled. Previously granted trials keep their expiry date.
Starter
For exploring and prototyping
charged as $15,000 CLP
- 1 vCPU
- 8 GB RAM
- 50 GB persistent disk
- Up to 3 saved workspaces
Pro
For serious projects
charged as $39,000 CLP
- 2 vCPU
- 16 GB RAM
- 50 GB disk
- Up to 6 saved workspaces
Team
For data teams
charged as $99,000 CLP
- 4 vCPU
- 32 GB RAM
- 100 GB disk
- Up to 12 saved workspaces
Every plan includes CPU and RAM reserved for your size and persistent storage. Plan changes are subject to available capacity.
Prices are shown in US dollars at a reference rate; the subscription is charged monthly in Chilean pesos, which is the amount next to each price and the one Stripe shows at checkout. Pausing a workspace does not cancel the subscription. Aquedra wallet usage, where applicable, is hourly. See Billing & refunds.
Questions before you start
What happens when an existing trial ends?
Your workspace stops and your data is preserved. Upgrade whenever you want and pick up exactly where you left off. If you never come back, we archive the workspace 60 days after expiry — we email you first, and the copy stays restorable on request for 180 days. Nothing is deleted behind your back.
Is my data private?
Every workspace is an isolated, unprivileged container with its own disk. No other user can reach your files, and we don't train anything on your data.
Can I take my work with me?
Yes. It's standard Jupyter and standard files: download your notebooks and data, or push to git from the built-in terminal, whenever you want.
What's preinstalled?
The Jupyter Docker Stacks images: Python with the SciPy stack (numpy, pandas, matplotlib, scikit-learn), R with tidyverse, a geospatial image and PyTorch. pip and conda work normally, and installs persist.
How does billing work?
Direct accounts purchase monthly CLP subscriptions through Stripe, renewing until canceled from Account. Pausing or deleting a workspace does not cancel the subscription. Aquedra wallet usage is separate and billed for running hours; these hourly charges are not added to a monthly subscription. The free trial requires no card and has no automatic charge.
What support do I get?
Real humans in English and Spanish. Write to [email protected] and a person answers — no first-line bot.
Built as a technical product
Real infrastructure, not a demo.
Real Kubernetes
Each notebook has its own CPU and RAM limits and reservations in Kubernetes. Resources are checked before creation.
Persistent storage + daily backup
Your disk lives on a persistent volume. Pause without losing data, and a daily job backs the volume up to object storage.
Per-user isolation
Unprivileged pods, no root, no Docker inside. Every resource is labeled and ownership-checked.
Support in English and Spanish
Written by the people who actually run the cluster. Ask about a stuck kernel or a custom image and you get a real answer.
Reserved resources
Your plan's CPU and RAM are accounted for as reservations. We distinguish allocated capacity from instantaneous usage.
Persistent disk + daily backup
Your volume survives pauses and restarts, and is backed up daily to object storage.
Our own infrastructure
Runs on our Kubernetes cluster in Europe (Hetzner) — not resold on top of someone else's notebook service.
Operated by Aquedra SpA
A real company behind it, with human support in English and Spanish.
Your next project starts here
Create your account and choose a plan with available stock to get started.
Plans subject to stock · cancel anytime · your data is kept