Parquet Reader and Viewer
Drop in a .parquet file and instantly browse its contents as a readable table, with column headers and inferred data types. Parquet is a compact, columnar format used by data pipelines, analytics engines, and machine learning tools, but it isn't human-readable on its own - this viewer decodes it locally so you can inspect a dataset without spinning up Spark, DuckDB, or a Python notebook.
Reviewed by Vincenzo Manto, Founder & Lead Developer · Last updated September 27, 2026
How it works
- 1
Drop a .parquet file into the upload area or select one from your device.
- 2
The file is parsed locally using hyparquet, entirely in your browser.
- 3
Rows render as an HTML table with column headers and inferred types.
- 4
For large files, the first 500 rows are shown with a note on the total row count.
Why view Parquet files in the browser?
Parquet's columnar layout makes it fast for analytics engines to query, but that same layout makes the raw bytes meaningless to a text editor. Historically, checking what's inside a .parquet file has meant installing Python, pandas, and pyarrow just to run a five-line script. This tool skips all of that: pick a file and see the table immediately.
Built on hyparquet, not a server
This viewer uses hyparquet, a dependency-free JavaScript parquet reader, to decode the file's metadata and column chunks directly in your tab. It works offline once loaded, handles files up to whatever size your device's memory allows, and never sends your data to a third party.
Frequently asked questions
Is my parquet file uploaded to a server?
No. The file is read in your browser using the hyparquet JavaScript library. Nothing is sent over the network, so it's fine for sensitive or proprietary datasets.
What compression codecs are supported?
Uncompressed and Snappy-compressed column chunks are supported out of the box, which covers the vast majority of parquet files produced by common tools like pandas, Spark, and DuckDB.
How many rows can I view?
The table shows up to the first 500 rows for performance. The summary above the table always shows the true total row count from the file's metadata.
Does it show nested or list/struct columns?
Yes. Nested values are flattened into a readable JSON-like string in their cell so you can still inspect the structure, though deeply nested schemas are easier to explore with the dedicated Schema Reader tool.
Why is my column showing as a bigint or hex string?
64-bit integer columns are decoded as JavaScript bigint (shown as plain numbers), and raw binary columns without a UTF8 annotation are shown as hex so no data is lost in the display.