macOS 15+ · Version 0.1.0

Open data.
Skip the stack.

Quick Look, browse, query, and inspect local data files in a native Mac app.

Parkour turns Parquet, CSV, JSON, databases, workbooks, Arrow files, and Markdown into first-class Mac documents—without uploading them or assembling a runtime.

Download DMG14 MB Download ZIP14 MB

Quick LookNative tableParquet inspectorLocal only

types.parquet

A document viewer for structured data

Your files should not require a notebook, a terminal, or an upload just to answer: what is in here?

Finder Quick Look

Press space. See the table.

Parkour adds native previews and branded thumbnails where Finder normally shows a generic document. The preview stays deliberately bounded so a glance remains fast.

Preview real rows with row, column, and format context
Works from Finder, the preview pane, and the spacebar
Slim readers keep the Quick Look extensions independent of the full data engine

Parkour Quick Look preview showing rows and columns from a Parquet file

Native browser

Millions of rows. Still a Mac app.

The main window pages through large files instead of loading them all at once. Sorting, filtering, column controls, copy, and export remain native controls—not a web table wrapped in a window.

Windowed scrolling with bounded, cancellable background queries
Sort, filter, resize, reorder, show, and hide columns
Export the shaped view to CSV or Parquet; run read-only SQL when needed

Parkour native table browser displaying typed Parquet columns

Parquet inspector

See what the file is doing.

Open the Inspector to understand a Parquet file without reaching for a notebook: compression, encodings, row groups, footer statistics, schema, and key-value metadata live beside the data.

Compression and encoding distributions by column and row group
Footer min/max, null density, and dictionary use without scanning data pages
Optional column summaries expose ranges, quartiles, and distinct counts

Parkour Parquet inspector showing compression metrics and column statistics

One app, many shapes

Open the file you have.

The same browser, shaping controls, SQL surface, and export path work across tabular formats. Format-specific detail appears only where it is real.

Parquet

Rows, nested types, metadata, row groups, codecs, and encodings

CSV + TSV

Dialect and type detection, including gzip and zstd streams

JSON + NDJSON

Arrays or streams, including gzip and zstd variants

SQLite + DuckDB

Browse each table through the same paging and query surface

Excel

Open .xlsx workbooks by sheet with typed cells and dates

Arrow + Feather

Open IPC files through Parkour's local data plane

Partitioned folders

Treat directories of matching part-files as one table

Markdown

Native rendered, source, HTML, plain, and prettified views

Built for local work

Private by architecture.

Local files stay local

The app contains no network client, analytics SDK, or cloud ingestion path. Opening a file does not upload it.

One self-contained app

The data engine and format support ship inside Parkour. Users do not need Homebrew, Python, a notebook, or a server.

Bounded by default

Large reads are paged, Quick Look previews are capped, and expensive summaries remain explicit on large files.

Get Parkour

Make data files feel native.

macOS 15+. Includes the app and its Quick Look preview and thumbnail extensions.