Convert a Streamlit app to marimo with AI
Move widgets, layout and calculations into a reactive Python notebook. Review how state, caching and custom controls behave after the conversion.
To convert a Streamlit app to marimo with AI, give streamlit-to-marimo the app and a short list of interactions readers depend on. Changing a date range might update both a table and a chart. A notebook that displays the initial figures correctly still needs to update both outputs when that date changes.
The skill maps widgets and layout concepts while taking account of the different execution models. In marimo, dependent cells react to changes in values. The old app's handling of state and caching therefore deserves an explicit review. Some custom Streamlit components need another implementation, which is worth identifying before a conversion is treated as complete.
What it does
- Move common controls into reactive notebook cells.
- Connect filter values to the outputs that depend on them.
- Recreate columns, sidebars and other layout elements.
- Review state management and cached calculations.
- Flag custom components without a direct equivalent.
Install and use
- Install the skill in a compatible agent with access to the app source. Use Python, marimo and the required project dependencies to validate the conversion.
- Ask “Migrate the sales dashboard. Switching the date range must change both the table and chart. Flag any custom widget that needs separate implementation.”
- Test several familiar inputs and compare the resulting behaviour with the original app.
FAQ
Source
Made by marimo. Original skill: streamlit-to-marimo, published on GitHub under the Apache-2.0 licence.
View on GitHub