Multi-user Jupyter, without operating a JupyterHub
JupyterHub is the standard way to serve notebooks to a class or a data team — and running it well means operating Kubernetes, Helm charts, storage, TLS, upgrades and auth. NotebookHub delivers the same outcome as a service: every user gets an isolated, persistent JupyterLab (or VS Code) workspace, and nobody on your team has to babysit a cluster.
What operating JupyterHub actually involves
The Zero-to-JupyterHub guide is excellent — and long, because the job is real: you'll provision a Kubernetes cluster, tune the Helm release, configure an authenticator and user storage, plan image updates, watch resource limits, and be on call when a student's server won't spawn the night before an assignment is due.
- A Kubernetes cluster to provision, patch and pay for
- Helm/z2jh upgrades and image builds to maintain
- Storage, TLS certificates and authentication to wire together
- Capacity planning so one heavy user doesn't starve the rest
What you get with NotebookHub instead
The same multi-user model — isolated per-user environments on Kubernetes with guaranteed resources — already operated for you:
- Per-user workspaces with reserved CPU/RAM (Kubernetes QoS Guaranteed) and persistent disks
- JupyterLab and VS Code in the browser, kernels for Python, R, geospatial and PyTorch
- Admin panel to manage users and plans
- SSO integration available for platform accounts
- Backups, upgrades, TLS and monitoring: our problem, not yours
For courses, research groups and data teams
Teaching a semester course? Every student signs up free and gets an identical environment — no laptop support sessions. Running a research group or a small data team? Paid plans add larger workspaces and more saved environments per user, billed simply, without per-cluster cloud invoices.
Need something specific — dedicated capacity, custom images, invoicing? Write to [email protected] and we'll set it up.
Frequently asked questions
Is this JupyterHub under the hood?
No — NotebookHub is an independent control plane on Kubernetes that provisions an isolated JupyterLab or VS Code container per user. Same outcome as a well-run Hub deployment, without you operating it.
Can every student really use it for free?
Yes — the free plan (one persistent workspace per account) is enough for most coursework. Group plans with bigger resources are available for departments.
Do users share resources?
No. Each workspace has reserved CPU and RAM — a heavy notebook in one workspace cannot slow down another.
Can we use our own authentication?
SSO is available through the Aquedra Cloud platform today; if you need a specific identity provider for your institution, contact us.
Serve notebooks to your team this afternoon
No cluster to build. First workspace free.