streamlit

Build data web apps from Python scripts

TLDR

Run Streamlit app

$ streamlit run [app.py]

Run with specific port

$ streamlit run [app.py] --server.port [8501]

Run with arguments

$ streamlit run [app.py] -- [--arg1] [--arg2]

Show configuration

$ streamlit config show

Create hello app

$ streamlit hello

Clear cache

$ streamlit cache clear

Show version

$ streamlit version

SYNOPSIS

streamlit [run] [config] [cache] [options] [file] [-- args]

streamlit creates web applications from Python scripts. Data apps, dashboards, and ML demos are built without web development knowledge.
Apps are Python scripts using Streamlit's API. Widgets like sliders, buttons, and text inputs are added with single function calls.
Hot reload automatically updates the app when code changes. The development workflow is fast and interactive.
Caching decorators prevent recomputation of expensive operations. Data loads and model inference are cached across reruns.
Deployment options include Streamlit Community Cloud, which hosts apps from GitHub for free. Self-hosting is also supported.
Components extend functionality. Custom widgets, visualizations, and integrations are available from the ecosystem.

PARAMETERS

run FILE

Run Streamlit app.

hello

Run demo app.

config show

Show configuration.

cache clear

Clear cache.

version

Show version.

--server.port PORT

Server port.

--server.address ADDR

Server address.

--server.headless BOOL

Run without browser.

--browser.serverAddress ADDR

Browser server address.

--theme.base THEME

Theme (light/dark).

--global.developmentMode BOOL

Development mode.

CAVEATS

Not suited for complex web apps. Limited layout control. Session state needs explicit management. Large apps may have performance issues.

HISTORY

Streamlit was created by Adrien Treuille, Thiago Teixeira, and Amanda Kelly in 2019. It rapidly gained popularity for data science prototyping and was acquired by Snowflake in 2022.