LOCAL DATA CONVERTER / NDJSON TO CSV

NDJSON to CSV

Turn flat NDJSON objects into a formula-protected UTF-8 CSV using a stable union of properties as columns.

Processed in this browserOne UTF-8 .ndjson or .jsonl file up to 5 MB, 50,000 non-empty lines, 200 union columns, 1,000,000 cells, and 20 MB output.

Step 1 / source

Choose the source

NDJSON - flat objects only

Drop one file here

Choose the flat ndjson objects you want to turn into utf-8 csv table.

Browse filesChecking publication status...

No file upload

This operation runs with browser APIs. Nirmion receives catalogue requests, not the selected source bytes.

Original stays unchanged

The converter creates a separately named output. It never writes back into the source file.

Bounded work

One UTF-8 .ndjson or .jsonl file up to 5 MB, 50,000 non-empty lines, 200 union columns, 1,000,000 cells, and 20 MB output.

Plain-language method

How NDJSON to CSV works

The browser validates each line independently as one flat object, collects property names in first-seen order, maps missing and null properties to empty cells, protects formula-like text when enabled, and quotes CSV-sensitive content.

01
Choose a valid source
Every non-empty line must contain one object with scalar values. Arrays, scalar-only lines, and nested values cannot form a predictable CSV table.
02
Review the conversion choices
Skip tolerates visual gaps; Reject enforces a strict stream. Formula protection is recommended for spreadsheet users.
03
Inspect the separate result
The CSV begins with union property names and contains one row per accepted NDJSON object.

Useful when

Practical jobs this tool handles

Review line-oriented records in a spreadsheet.
Prepare flat logs for a CSV importer.
Audit union columns across varying objects.

Before converting

Check the source and destination rules

Keep one complete object per line.
Flatten nested values before conversion.
Review the union header and formula protection.

Know the boundary

Know what the destination format cannot keep

CSV loses the distinction between null and missing properties, and it cannot represent nested objects or arrays. Those inputs are rejected rather than stringified ambiguously.

Quick answers

Questions before you convert

The conversion contract is deliberately explicit.

Can property order vary between lines?

Yes. Columns follow the first appearance of each property.

Are malformed lines skipped?

No. One malformed non-empty line stops output so record loss is visible.

Can a line be a JSON array?

No. This route requires flat objects for a valid table.