Products
JupyterHub
Managed GPU notebooks for a whole class or lab. Everyone gets a real slice of an H200, nobody has to administer a server, and notebooks left idle stop billing.
Early access is open to enterprises on NVIDIA H200. More GPUs, and access for more teams, are coming soon. We reply within one working day.
built for universities, bootcamps and r&d teams. Technology: multi-user · spawner profiles · idle culling
Pick-a-size profiles
Offer a CPU-only profile for data prep and H200 slices for training. Users choose from a list; they never see a node or a scheduler.
Idle notebooks shut down
Notebooks left open overnight stop billing. You set the idle limit per organisation; by default a server stops after 60 idle minutes.
Fair limits per person
Cap servers, memory and GPU share per user, so one runaway notebook can't use up a department's quota.
Shared and personal storage
A private volume per user plus a shared, read-only dataset mount, so a class of forty works from one copy of the data, not forty.
Course-ready access
Invite a whole cohort at once, hand out a profile, and remove access when term ends. No cloud accounts, no credit cards, no foreign billing.
Know where the budget went
Every notebook-hour is metered to the user and the project, so a department can see exactly where its money went.
- Hub version
- JupyterHub 5.2
- Profiles
- CPU · H200 MIG 1g · MIG 2g
- Billing
- per user-hour, metered per second
- Idle shutdown
- configurable, 60 min default
Start with a quote.
Tell us what you want to run. We reply within one working day with a capacity plan and a firm rupee quote.
Other products