Data Workbench

Open CSV, JSON, Parquet, and Excel files. Inspect the data, run SQL queries, and convert between formats in your browser.

Built with DuckDB-WASM.

Choose a file or paste a URL to get started. You can inspect the data, query it with SQL, or export it to another format.

Loading remote files or database extensions makes requests to those services.
For best experience, use a tablet or larger screen

Import Data

Choose a data file or paste a URL.
You can also drop files here.
Supports: CSV, JSON, JSONL, Parquet, Excel

Getting the database ready…

When enabled, DuckDB will skip problematic rows and load valid data. Useful for CSV files with inconsistent formatting or Excel files with mixed data types.
CORS Notice: Some URLs may fail due to cross-origin restrictions. Services like HuggingFace, AWS S3, and Google Cloud Storage typically work well. GitHub raw files and many other services block cross-origin requests.
For authenticated URLs: Open the full interface and use the SQL Console to set up credentials first (e.g., SET s3_access_key_id='...';). Check the DuckDB documentation for specific authentication settings for your cloud provider.

Inspect a data file, run SQL, or change formats

Open a dataset without setting up a database. DuckDB runs inside your browser.

  1. Choose a local file or enter a URL. Inspect the preview and detected columns.
  2. Use the SQL Console to query the loaded input_data table.
  3. Use the export controls to save data in another available format.

Preview a CSV before converting it to JSON

Load a CSV, run this query to check a few rows, then choose JSON in the export controls. If the columns look wrong, fix the import before exporting the whole thing.

SELECT *
FROM input_data
LIMIT 10;

What to know before you start

The workbench handles CSV, JSON, newline-delimited JSON, and Parquet. Excel .xlsx support depends on the DuckDB Excel extension loading successfully.

Local files are processed in the browser. Loading a URL contacts that host, and downloading the engine or extensions contacts their providers. Some remote files cannot be opened because their host restricts browser access.

Available memory limits dataset size. Converting a file does not preserve spreadsheet formatting, formulas, or workbook layout.