dataeng.tools

Open a data file from a URL

Paste a link to a Parquet, Avro, CSV, TSV, JSON or JSONL file — a public S3 or GCS object, or any https URL — and query it directly with DuckDB, no download step. Your browser fetches it (over HTTP range requests for Parquet) and reads it locally; nothing is uploaded to us.

Paste a public link to a Parquet, Avro, CSV, TSV, JSON or JSONL file. It’s fetched by your browser (the host must allow CORS) and read locally — nothing is sent to us.

Frequently asked questions

Which URLs work?
Any publicly-readable https URL, including S3 and GCS objects that allow anonymous or presigned access. Your browser fetches the file directly — we never see it.
Does it download the whole file?
For Parquet it uses HTTP range requests to read only the metadata and the columns/rows you query, so you can open large remote files without pulling them in full. CSV and JSON are fetched completely.
Can I open a private S3 or GCS object?
Yes, if the URL is presigned or the bucket allows anonymous reads. The request goes straight from your browser to the store; no credentials pass through us.
Is anything uploaded?
No. Your browser fetches the URL and reads it locally with DuckDB-WASM; the data never passes through our servers.

Related tools