Jupyter with GPU, ready to use
Launch your environment in minutes and install what you need.
Ready-to-use environments for analysis, notebooks and experimentation
Datoki is for people and teams who need to open Jupyter with GPU without preparing a machine from scratch every time.
It relies on distributed infrastructure to make available capacity usable for real work without forcing you to build the technical layer yourself.
Three simple steps
Sign in
Sign in to see your balance and get started.
Use your balance
You decide when to spend and how much room you have before launch.
Launch your environment
Open the environment you need and start working.
Sign in to choose your environment
Once inside you can see your balance and launch the environment that fits your work.
Choose the environment you need
No subscription. You pay only for the session you launch, when you launch it.
Explore
For exploratory analysis, notebooks and first tests. Ideal if you are just getting started or do not need more power.
VRAM
12 GB VRAM
Duration
60 min
If you are not sure which one to choose, this is usually the best place to start.
Train
For training models and experiments that need more muscle. When Explore starts to feel too small.
VRAM
24 GB VRAM
Duration
90 min
A good fit when Explore starts to feel too small, but you still do not need the most powerful machine.
Research
For heavy workloads, large models and serious work. When you already know what you need.
VRAM
40 GB VRAM
Duration
120 min
Choose this when you already know you will be working with larger loads or want the most headroom available.
Beta access
Leave your email and we will review your access for this beta round. If there is room, you will receive an individual single-use code to try Datoki.
Requests are reviewed manually from Admin and access may close once the quota is exhausted.