Skip to content
OnDevKit

DuckDB SQL Explorer

Run SQL on CSV, Parquet and JSON files with DuckDB, right in your browser. Files are never uploaded.

  • Never uploaded
  • Works offline
  • Free, no sign-up

Data and query

DuckDB · runs in this tab
CSV, TSV, Parquet or JSON. Drop files anywhere here. They’re read on this device.

Results

Add a file to query it with SQL, or try it with 600 sample orders.

Processed on this device. Nothing is sent to a server.

How to use the DuckDB SQL Explorer

  1. Step 1Add your filesDrop CSV, TSV, Parquet or JSON files. Each becomes a table.
  2. Step 2Write SQLJoins, window functions, GROUP BY. Press Ctrl+Enter to run.
  3. Step 3Export the resultsCopy or download them as CSV.

SQL on your files, without a database server

DuckDB is an analytics database built for exactly this: reading files directly and answering SQL queries over millions of rows in seconds. Here it runs as WebAssembly inside your browser tab, so you get the full SQL dialect, including joins across files, window functions, PIVOT, regular expressions and date functions, without installing anything.

Files are read from your disk by the browser and never leave your device. That makes it safe for customer exports, financial data and logs you can’t upload to an online converter. Column types are detected automatically, and SUMMARIZE gives you min, max, null counts and distinct values for every column at once.

Questions

Are my files uploaded?

No. DuckDB runs in your browser, and files are read directly from your device. Even the DuckDB engine and its Parquet and JSON add-ons are served from this site, so the tool contacts no third party.

How large can my files be?

Files of a few hundred megabytes work well on most computers. The limit is your browser’s memory. Parquet files are far more compact than CSV, so they go further.

Why does the first query take a few seconds?

The DuckDB engine (about 8 MB) downloads the first time you use the tool. After that your browser caches it, and it works offline.

Can I join two files?

Yes. Each file becomes a table named after the file, so you can write SELECT … FROM orders JOIN customers USING (customer_id).

Where can I learn DuckDB’s SQL?

DuckDB follows PostgreSQL closely and adds conveniences like GROUP BY ALL, SELECT * EXCLUDE (…) and SUMMARIZE. The DuckDB documentation at duckdb.org lists every function.

Related tools