Convert Parquet to JSON Online — In-Browser, SQL-First

Parquet is excellent for storage and analytics. JSON is what everything else speaks. PondPilot converts Parquet to JSON in the browser — and lets you shape the output with SQL so you can produce something a human or an API will actually accept.

How to Convert

  1. Open app.pondpilot.io
  2. Drop your .parquet file
  3. Write a query (or SELECT * FROM file.parquet)
  4. Export as JSON

The whole pipeline runs locally via DuckDB WebAssembly. No server, no upload.

Shape for APIs or Readability

A blind Parquet-to-JSON dump is rarely what you need. SQL lets you pick columns, rename fields, and build nested objects:

SELECT
  order_id AS id,
  {'name': customer_name, 'email': customer_email} AS customer,
  total::DECIMAL(12,2) AS total,
  order_date::DATE AS date
FROM 'orders.parquet'
WHERE order_date >= '2024-01-01'

Export the result as a JSON array or NDJSON. Good for API fixtures, debugging payloads, or sharing a readable slice of a larger dataset.

Big Parquet, Small JSON

Parquet files can be huge. JSON of the whole file would be enormous and unreadable. The point of converting to JSON is usually a subset — a sample, a filter, a specific account. DuckDB’s columnar reader only pulls the data your query touches, so extracting a slice from a big Parquet is fast.

Inspect First

Not sure what’s in the file?

DESCRIBE SELECT * FROM 'unknown.parquet';
SELECT * FROM 'unknown.parquet' LIMIT 5;

See the schema, preview some rows, then write the query that produces the JSON you want.

Privacy

Parquet files often hold production data, customer records, or analytics events. PondPilot processes everything in-browser. The file stays on your machine.

Get Started

Visit app.pondpilot.io and convert your Parquet to JSON.

Back to home