Streamlit

Streamlit is a Python framework for building data apps, dashboards and internal tools without writing HTML, CSS or JavaScript. It is aimed at people who already have Python code, such as data analysis or a model, and want to put a UI in front of it quickly. Streamlit is owned by Snowflake.

How it works: you write a normal Python script that calls functions like st.write, st.slider or st.dataframe, and run it with streamlit run. Widgets behave like variables: when a user moves a slider, Streamlit reruns the whole script from top to bottom with the new value. @st.cache_data lets expensive steps skip work on reruns, and st.session_state keeps values between them.

What you get: a library of widgets, charts and layout elements; multipage apps via st.navigation and st.Page or a pages/ folder; theming through .streamlit/config.toml; user login through OpenID Connect providers with st.login and st.user; and custom components when you need a widget it doesn't ship, built with HTML, CSS and JavaScript. You can self-host it (the docs cover Docker and Kubernetes), deploy free on Streamlit Community Cloud from a GitHub repo, or run it inside Snowflake.

The limits: the rerun-everything model is simple, but slow code runs again on every interaction unless you cache it, and the look is largely what the built-in elements and theme settings allow. On Community Cloud, apps inherit visibility from their GitHub repo, you get one private app at a time, and apps sleep after 12 hours without traffic.