We connect your Jupyter notebooks to any compute: ours, yours, or someone else's!

Turbocharge your research

ML experiments often start as notebooks on your local machine, but grow into computationally intensive workflows that require more powerful compute.

Right now, this means spending time on setting up remote GPUs and dealing with all the messy configuration which comes with that (e.g. SSHing into them, doing package management, pulling your repository and loading the data you need).

Instead, we let you keep the same Jupyter interface in your local IDE that you know and love. Just click a button to change the Python kernel from "local" to "remote", and you'll be connected to your compute provider.


We support connecting notebooks in VS Code / Cursor to Runpod instances. If you want to use Moonglow with a setup we don't yet support, let us know!

Connecting to AWS, Azure, GCP etc.
Coming in July!
Connecting to multi-node clusters
Coming in August!
Connecting to custom clusters on prem
Contact us!

Who are we?

We are a team of researchers and engineers with a history of working on ML research and building high-performance infrastructure.

Backed by
© 2024 Moonglow
Making great software for members of technical staff!