Convert CSV to a JSON array or JSON Lines: any delimiter, with or without a header row, typed numbers, zip codes kept as text
Turn a CSV or spreadsheet export into a JSON array of objects, using the first row as keys.
How to use it
Paste CSV or open a .csv file. With Header row on, the first row names the keys; off, each row becomes an array. Pick the delimiter (comma, semicolon, tab, pipe, or Other for any 1 to 4 characters). Under Options: the quote character, lines to skip as comments, JSON Lines, and the indent or Minified. Choose whether numbers and true/false become typed values, then copy or download the JSON.
Examples
Typed values and nesting
Input
id,user.name,zip,active 1,Ada,02134,true
Result
[
{
"id": 1,
"user": {
"name": "Ada"
},
"zip": "02134",
"active": true
}
]Leading zeros are kept as text, so zip codes and IDs are not damaged.
No header row: arrays
Input
1,Ada,true 2,Grace,false
Result
[[1,"Ada",true],[2,"Grace",false]]
Header row off, Indent Minified.
JSON Lines, with a comment line and a ~ delimiter
Input
# exported 2026-11-10 id~name 1~Ada 2~Grace
Result
{"id":1,"name":"Ada"}
{"id":2,"name":"Grace"}Delimiter Other “~”, Comment lines “#”, Output JSON Lines: one object per line, as log tools and BigQuery load them.
Type detection rules
Plain decimal numbers like 42 or -3.5 become JSON numbers, with the digits the cell had: 1.10 stays 1.10. Numbers with leading zeros (02134), numbers with more digits than JavaScript holds, and exponents such as 1e5 stay strings. true and false become booleans; an empty cell becomes null. Everything else stays a string. Turn detection off to keep every value as text.
Questions
- Does the CSV need a header row?
No. With Header row on, the first row gives the keys: two columns with the same name become name and name_1, a short row gets empty values for the missing cells (null with Detect types on), and a row with more values than the header stops with its row number. Turn it off and every row, the first too, becomes an array of its values. - What is JSON Lines?
One JSON value per line, each line ending in a line break, with no array around them (also called NDJSON). Log tools, jq and data loaders such as BigQuery read it a line at a time, so a big file needn’t fit in memory at once. - Does it handle quoted fields with commas and line breaks?
Yes. Standard CSV quoting is supported: "Hopper, Grace" stays one value, and doubled quotes inside a field are unescaped. For files quoted with single quotes, pick Quote: Single under Options. - Can I get nested JSON from a flat CSV?
Yes. Keep "Nest dotted columns" on and a column named address.city becomes {"address": {"city": …}}. - Is my data uploaded?
No. Everything runs in your browser, so your data stays on your device. Nothing you paste is uploaded unless you share a result, which is locked with your password first. That matters for exports with customer or user data.