Streamlit apps: Deploy Instantly with Zero Configuration

I’ve written here before about a proof-of-concept for deploying Streamlit inside JupyterHub.

I’m excited to share that these ideas have now made it into the ContainDS Dashboards extension for JupyterHub, so you can automatically deploy a Streamlit app accesible to all colleagues within a JupyterHub. You only need to upload the Streamlit script to your home folder, then specify the script name to register it as a new Dashboard.

This means you can also continue to use your Jupyter notebooks as normal, and don’t have to fiddle with hosting and authentication just to share your Streamlit work with colleagues!

More details on this simple workflow are here:

10 Likes

Really interesting @danlester - thanks for sharing!

1 Like

Update: this can now work with Kubernetes-based JupyterHubs (Zero to JupyterHub).

This means a user can upload Streamlit scripts to their usual Jupyter server, then instantly convert them into a new user-friendly server to share with non-technical colleagues who are authenticated by the organization’s JupyterHub.

This also works for Plotly Dash and Jupyter notebooks (displayed as Voila apps).

Kubernetes support is configuration-dependent, so please get in touch to talk through your setup - I’d be really pleased to help. K8s installation is discussed here.

2 Likes

Hi @danlester.

Thats great. Then you can actually deploy using Streamlit, Panel, Voila and Dash.

Awesome.

3 Likes

Definitely - let me know how you get on with it!

Thanks,

Dan