Is my Parquet file uploaded anywhere?

No. The file is decoded and converted entirely on your device, in the browser tab. Its bytes are never sent to a server, so even sensitive datasets stay with you.

What is a Parquet file?

Parquet is a compact, binary, columnar storage format used across the big-data and analytics world. It is built for engines to read quickly, not for people to open and read by eye.

What does the conversion produce?

A single Markdown pipe table. The column names come from the keys of the first row, and every following row becomes a table row, so the whole dataset reads as plain text.

Are the column types kept?

No. Markdown tables have no schema, so every value is written as text. The numbers, booleans and dates are all correct, but the column is no longer typed as an integer or a timestamp.

How are nested values handled?

Lists and nested records are written as JSON text inside the cell, for example `["a","b"]` or `{"k":1}`. The value is preserved faithfully, just not as a nested structure.

What happens to dates and timestamps?

Dates are written as ISO 8601 text, such as `2024-01-05T00:00:00.000Z`. Large whole numbers stored as bigints are written out in full as their digits.

Can it handle very large Parquet files?

It can, but be aware that the whole file is read into memory to decode it; there is no streaming. A very large dataset can therefore be heavy on a low-memory device, so expect it to take a moment.

Is the output good for LLMs and RAG?

Yes. A flat Markdown table is easy for a model to read, and you can add YAML front matter, a table of contents or export RAG chunks from the result panel.

Convert Parquet to Markdown.

Drop a Parquet file and read its rows and columns as a plain Markdown table. It runs entirely in your browser, so your file never leaves your device.

Instant & offlineFree, no accounts

Drag & drop your files

WordExcelPowerPointPDFEPUBCSVJSONCode

+135 file formats supported

Batch convertMany files at once

100% privateStays on your device

Works offlineNo connection needed

Preset

Optimize for AI & RAG

Extra cleanup for LLM ingestion: strip HTML, fix smart quotes, tidy Unicode and spacing.

Add YAML front matter

Prepend a metadata block (title, source, date, word & token counts) for knowledge bases and RAG.

Add table of contents

Build a linked index from the headings. Handy for long documents.

Export RAG chunks (.json)

Split the result into retrieval-ready chunks. Download per file from the result panel.

Most converters quietly upload your documents to a server. This one physically can't.

01WhyPrivate Markdown

Built for engines,
not for eyes.

Parquet packs your data into a compact binary you cannot open and read. Converting decodes those columns and lays every record out as a readable Markdown table.

PARQUETdata.parquet

columnarbinary

Everything you
actually need.

Columnar data in, a clean Markdown table out, with no server and no account anywhere.

It never leaves your browser

The Parquet file is decoded and converted on your own device. Nothing is uploaded to any server, ever.

local

PARQUETdata.parquet

| id | name |

| 1 | rows |

3 chunks

AI & RAG ready

A flat table feeds a model cleanly, with optional front matter and RAG chunk export.

Works offline

Once the page has loaded you can switch off your connection and it keeps converting.

Reads the binary

Parquet is a compact binary, columnar format; it is decoded right in your browser and unrolled into table rows.

éñü✓

Unicode safe

Text values, accents and non-Latin scripts come through intact as UTF-8.

Free, and unlimited

No sign-up, no quotas, no watermarks. Convert one dataset or a thousand; it all runs the same way, on your own device.

03FidelityPrivate Markdown

Values in,
types out.

Every cell makes it across, but a column is no longer typed as a number or a date. These are the same notes the Formats list shows for Parquet, so the page never drifts from what the converter really does.

Kept

Every row
Every column
Header row
Cell values

Dropped

Column types
Nested structure
Columnar layout

PARQUETdata.parquet

kept

int64stringtimestamp

dropped

{ "tags": [ "a", "b" ] }

dropped

04FAQPrivate Markdown

Parquet questions,
answered.

Everything worth knowing before you drop in a Parquet file.

05MorePrivate Markdown

Other converters.

Working with more than Parquet? These convert the same way: privately, in your browser.

SAMI to Markdown

.smi · .sami

Microsoft SAMI captions.

Convert

TTML to Markdown

.ttml · .dfxp

Timed-text captions.

Convert

SBV to Markdown

.sbv

YouTube SBV subtitles.

Convert

SubViewer to Markdown

.sub

SubViewer subtitles.

Convert

GeoJSON to Markdown

.geojson

Geographic JSON features.

Convert

GPX to Markdown

.gpx

GPS tracks & waypoints.

Convert

KML / KMZ to Markdown

.kml · .kmz

Google Earth geo data.

Convert

TopoJSON to Markdown

.topojson

Topology-aware GeoJSON.

Convert

Browse all +135 formats

Convert Parquet to Markdown.

Drag & drop your files

Built for engines,not for eyes.

Everything youactually need.

It never leaves your browser

AI & RAG ready

Works offline

Reads the binary

Unicode safe

Free, and unlimited

Values in,types out.

Parquet questions,answered.

Other converters.

Built for engines,
not for eyes.

Everything you
actually need.

Values in,
types out.

Parquet questions,
answered.