Open Parquet files in a spreadsheet

A native Parquet viewer for macOS, Windows, and Linux.

Typed columns, not guessed text.

Free and open source.

Parquet is what data pipelines write, but opening one usually means a Python one-liner or a query tool. VisiGrid opens a .parquet file like any spreadsheet: look through it, run formulas on it, and export it to CSV or Excel.

See the data without writing code.

macOS · Windows · Linux · vgrid convert

A Parquet file is not something you can just open

Parquet is a compressed, columnar, binary format. That is exactly why pipelines use it, and exactly why double-clicking one gets you nowhere.

The usual answers are a pandas or DuckDB one-liner, a notebook, a query editor, or uploading the file to a website — which is a poor trade when the data is a customer export or anything you would rather keep on your machine.

Often you only want to answer a simple question: did the job write what it should have, what is in this column, how many rows came through.

VisiGrid puts the file in a grid, with the types the file declared, in about the time it takes to open a spreadsheet.

Column types survive the trip

Parquet already knows its types, so nothing is guessed from text the way CSV import has to. A string column holding 007 stays 007.

Dates, times, and timestamps arrive as real dates. Decimals keep their exact value, and a decimal too precise for a spreadsheet number stays as text rather than being silently rounded.

Structs, lists, and maps do not fit in a cell, so they show as readable JSON text instead of disappearing.

Local, native, and offline

A Rust desktop app, under 30 MB, rendering on the GPU. No Electron, no browser tab, and no account.

The file never leaves your machine, and large imports run in the background so the window stays responsive.

Once it is open, it is a spreadsheet: formulas, formatting, and export to CSV, TSV, JSON, or Excel for whoever asked.

Parquet to CSV from the command line

The vgrid CLI uses the same reader, so a conversion can run in a script or on a server with no display.

vgrid convert events.parquet -t csv -o events.csv

Reading only: VisiGrid opens and converts Parquet, but it does not write Parquet files.

The limits, up front

A sheet holds 1,048,576 rows (including the header) and 16,384 columns.

Open a bigger file in the desktop app and it loads what fits, then tells you how much was left out. The CLI refuses a file that does not fit and suggests splitting it first.

For multi-million-row files, a query engine is the right tool. VisiGrid is for looking at, checking, and handing over the data that fits in a grid.

When it matters

Checking a pipeline output

Open what the job wrote and see the columns, types, and values before anything downstream reads it.

Handing data to someone else

Open the Parquet file and export it to .xlsx or CSV, with dates still readable and IDs still text.

A quick question about a file

Count, sum, or look something up with a formula instead of opening a notebook.

Data you cannot upload

Customer exports and internal data stay on your machine. No account, and it works offline.

Common questions

How do I open a Parquet file?
Download VisiGrid for macOS, Windows, or Linux (version 0.34.0 or later) and open the .parquet file the same way you open a spreadsheet. The column names become the header row and each record becomes a row.
Can I convert Parquet to CSV or Excel?
Yes. In the desktop app, open the file and export it to CSV, TSV, JSON, or .xlsx. From the command line, run vgrid convert data.parquet -t csv -o data.csv. Dates and timestamps are written as ISO 8601, such as 2026-09-01 14:02:00.
Is there a file size limit?
A sheet holds 1,048,576 rows including the header and 16,384 columns — the same as Excel. The desktop app loads what fits from a larger file and tells you what was left out. vgrid convert refuses a file that does not fit, so split or filter it first. A file that fits still has to fit in memory: a million rows of five columns needs roughly a gigabyte while it is open.
Does VisiGrid keep the data types from the Parquet file?
Yes. Numbers, booleans, dates, times, timestamps, decimals, and UUIDs keep the type the file declared. Text is never reinterpreted, so leading zeros survive. Decimals with more than 15 significant digits are kept as text so the value is not rounded.
What happens to nested columns?
Structs, lists, and maps are shown as JSON text in the cell, so they stay readable and can be parsed later.
Can I edit a Parquet file and save it back?
No. VisiGrid reads Parquet but does not write it. You can change the data in the sheet and export it to CSV, TSV, JSON, or .xlsx, or save it as a VisiGrid .sheet file.
Is there an online Parquet viewer?
The VisiGrid web app at app.visigrid.app imports Parquet files up to 30 MB when you are signed in. If the data should stay on your machine, use the desktop app, which needs no account and works offline.
Is VisiGrid free?
Yes. VisiGrid is free and open source under AGPLv3, including Parquet support. Paid plans add capabilities for larger and accountable work; see the pricing page.

Full reference, including how every Parquet type converts and the CLI commands: Open Parquet Files in the docs.

Working with CSV instead? See how VisiGrid compares to Excel, LibreOffice Calc, and the other usual tools.

Open your Parquet file

Free and open source (AGPLv3) for macOS, Windows, and Linux. Parquet support is in VisiGrid 0.34.0 and later.

vgrid convert data.parquet -t csv

Download Free See Pro features