datasette-marimo 0.2.0

datasette-marimo

Use marimo inside of datasette.

Install

Install this plugin in the same environment as Datasette.

uv pip install datasette-marimo
datasette install datasette-marimo

Demo

The demo on GitHub Pages shows the notebook experience on a Datasette server.

Usage

Run a Datasette server. Open the top-right hamburger menu and click Open in marimo. You can also open /marimo directly.

marimo then runs in WASM with a connection to your Datasette instance. You can run visualization and machine learning tools on the data. You do not install any software on your local machine.

Warning: if you refresh the page, you lose your work. Export your work before you refresh.

The Open in marimo link reads the page you came from. It prefills the database and table, so the notebook opens ready to query that table.

Inside the notebook you connect through a marimo SQL connection from moutils. The connection appears in marimo's data-sources panel. You can browse the schema and write native SQL cells against it:

from moutils.db.datasette import DatasetteConnection, databases

databases(base_url)                          # list databases on the instance
conn = DatasetteConnection(base_url, "sqlite")  # connect to one database

# then, in a SQL cell (or from Python):
mo.sql("select * from chickweight limit 100", engine=conn)

The moutils connection runs synchronous HTTP in the browser. It needs a browser with JSPI support, such as Chrome or Edge. On other browsers, use the legacy helper below.

The notebook still includes the older Datasette helper class (get_polars / sql_polars) for backward compatibility. DatasetteConnection is the recommended path.