Datasette

Datasette is an open source tool for exploring and publishing data, aimed at data journalists, museum curators, archivists, local governments, scientists and researchers. It turns data of any shape into an interactive, explorable website with an accompanying API. The project is hosted by simonw on GitHub under the Apache 2.0 license, and the latest stable release listed is 0.65.5, with 1.0 alpha releases also in progress.

What It Is

Datasette is a Python application that serves SQLite databases through a web interface and a JSON API. You point it at a database file, for example with datasette serve path/to/database.db, and it starts a local web server on port 8001 for browsing the tables. Data can be imported from CSVs, JSON, database connections and more.

Publishing and Prototyping

The datasette publish command packages one or more SQLite databases and the application into a Docker image and deploys it to hosting providers such as Heroku or Google Cloud Run. The homepage also lists Vercel as a publishing target. A metadata.json file adds title, license and source information, which appears on the index page and in the API output. The JSON API can also be used to prototype ideas without building a custom backend.

Plugins and Ecosystem

Datasette is part of an ecosystem the homepage describes as 44 tools and 154 plugins. Plugin hooks cover areas such as write wrappers, token handlers, background tasks and JavaScript modal dialogs. Datasette Agent, an open source plugin announced in May 2026, adds an extensible AI assistant for interacting with SQLite databases. Datasette Apps, announced in June 2026, hosts custom HTML applications inside a Datasette instance. Datasette Lite packages Datasette with WebAssembly so it runs entirely in the browser.

Recent Releases

The 1.0 alpha series has added stored queries and SQL write queries, a Jump menu, column types, a SQL-powered permissions system and OpenTelemetry traces and metrics in 1.0a41. Releases 0.65.5 and 1.0a40, from September 2026, include a security fix for a table permissions bypass.

Setup Path

It installs with Homebrew, pip or pipx, or with Docker, and requires Python 3.11 or higher.