100% client-side · 0 trackers · 0 uploads · verify in DevTools → Network
Uploading a spreadsheet just to run a query means handing a third party your raw data: customer rows, financials, anything in the file. data0 skips the server entirely. DuckDB-WASM loads once from a CDN, then every query runs locally over WebAssembly. Open the Network tab and you will see zero outbound requests while you work.
Drop a .csv, .tsv, .parquet, .json, or
.ndjson file and query it with full DuckDB SQL: CTEs, joins, window functions,
and aggregations all work. Files become tables named file1, file2, …
so you can join several at once. No account, no upload, no server. Press
Ctrl/⌘ + Enter to run a query.