QuackDB is a privacy-preserving in-browser DuckDB SQL playground and editor.

Demo available at quackdb.com.

• In-Browser: The DuckDB Wasm library is used to run SQL queries in the browser.
• File Types: Supports .csv, .json, .parquet, .sqlite, .duckdb and .arrow files.
• Privacy: no screen recordings or client-side error monitoring (only Vercel Analytics).
• Filesystem Access: OPFS is used to store files in the browser's filesystem for persistence.

I wanted a quick and simple tool for iterating on DuckDB queries and visualizing the results. I also wanted to experiment with the new Filesystem Access API (OPFS) and explore the capabilities of DuckDB in the browser.

An earlier version of this project used SST for the backend infrastructure, but I switched to Vercel for simplicity.

• Safari support 😁
• Improved UI / charts / tables / etc.
• Data import/export
• Responsive design improvements
• Performance improvements with Arrow
• Snippets / examples

• NSQL integration for natural language queries.
• Co-pilot (or similar) integration for code suggestions.

• CMU Data Interaction Group
• SQL Workbench
• Evidence
• Malloy
• wa-sqlite
• Rill
• Observable

MIT

This project is not in any way affiliated with DuckDB or any of the other projects mentioned above.

• Built in chrome OPFS viewer can be accessed from https://www.quackdb.com/temporary. To navigate, manually change the URL to the desired path as Chrome does not allow navigation by clicking on the links.

• Safari will not be fully supported until this bug is fixed (serializing file handles is not supported in Safari).