I’m the creator of dstack, a tool that makes it easier to train models in the cloud. Our tool allows extending it with custom providers to support different languages, frameworks, etc. Today, we’ve released a new update that extends the capabilities of dstack beyond training models, and now also allows users to quickly build and share apps with Streamlit (and other frameworks, incl. FastAPI) in the configured cloud – in just a few clicks.
All the built-in providers are also open-source: GitHub - dstackai/dstack: 🧬 dstack: Your new home for building AI apps
I invite everyone to give it a try and share their thoughts. Happy to discuss the approach and what would be great to have in terms of other features!
Blog post: Introducing apps and dev environments
P.S.: Currently, it’s possible to run models and apps only in the configured cloud. If you’d like the tool to also allow you to run it locally, and if you would like this part to be open-source too, please leave comments!