Skip to content
FizzKit

CSV to JSON

Convert CSV or TSV data into JSON. Auto-detects the delimiter, handles quoted fields, custom quote / escape / comment characters, and outputs an array of objects, array of arrays, column-oriented object or NDJSON. Type inference turns numbers and booleans into JSON values while preserving leading zeros and 16+ digit identifiers as strings, and dotted headers such as address.city can be expanded into nested objects.

Runs locally in your browser
For large CSVs, open fullscreen for a wider view
CSV input
0 chars · 1 line
JSON output
Output will appear here…
Delimiter
Indent

Auto-detects comma, semicolon, tab and pipe. Ragged rows are reported as warnings instead of failing the whole file. "Columns" outputs one array per header (column-oriented), "Arrays" keeps rows as arrays with the header as the first row.

How to use CSV to JSON

  1. 1Paste or upload CSV or TSV. The delimiter is detected automatically or can be forced.
  2. 2Check the Table preview for the first 100 rows and any parse warnings.
  3. 3Choose the output shape: Objects, Arrays, Columns or NDJSON.
  4. 4Adjust type inference, leading-zero preservation, trimming and nested header expansion, then copy or download.

About CSV to JSON

Convert CSV or TSV data into JSON. Auto-detects the delimiter, handles quoted fields, custom quote / escape / comment characters, and outputs an array of objects, array of arrays, column-oriented object or NDJSON. Type inference turns numbers and booleans into JSON values while preserving leading zeros and 16+ digit identifiers as strings, and dotted headers such as address.city can be expanded into nested objects.

CSV to JSON is part of FizzKit — a collection of focused developer and ad-tech tools. Because everything runs locally, it works offline once loaded and never exposes your data to a remote server.

Frequently asked questions

Yes, it chooses between comma, semicolon, tab and pipe based on the input, and shows which one it picked. You can force a delimiter if the guess is wrong.