CSV to JSON Converter

A naive conversion makes every value a string, so "qty": "3" arrives where a number was expected and a spreadsheet import quietly stores the wrong type. This tool reads each cell and converts plain integers, decimals, true, false and null to their JSON equivalents.

Two safeguards stop that from going too far. A value with leading zeros such as 007 stays text, because a product code or a ZIP is not a number. And anything with more than 15 digits is left as a string rather than being rounded by a floating point number that cannot hold it exactly.

Files exported from spreadsheets in different locales use different separators: a comma in English, a semicolon in much of Europe, a tab or pipe in older systems. The tool tries all four and keeps whichever produces the widest rows on average.

That method only works if quoted fields are respected, and this parser follows RFC 4180 while doing it. A field such as "Leeds, UK" containing a comma is one field, not two, and a doubled "" inside a quoted field becomes a literal quote. Get that wrong and one stray comma makes every later row look ragged, which is the single most common reason a CSV import produces a mess.

With the header box ticked, the first row supplies the field names and each later row becomes one object. Problems that would otherwise produce broken JSON are repaired rather than ignored: a repeated name becomes name_2, a third becomes name_3, and a blank column heading becomes column_4.

Untick the header box and each row becomes an array of values in the order the columns appear, which is what you want for a file that genuinely has no field names. Values are still typed either way.

If a data row has fewer cells than the header, the missing ones become empty strings so the JSON stays rectangular and every record has the same keys. The result also tells you how many rows were affected, because a row count mismatch is almost always a delimiter appearing unquoted inside a value and is worth fixing at the source.

Blank cells become empty strings rather than null. In tabular exports a blank cell usually means the field was left empty, not that the value is unknown, and the difference matters when the JSON feeds a database import.

The usual destinations are API request bodies, seed data for a local database, fixtures for a test suite, and bulk imports through tools that expect JSON Lines or a single array. Because the output is generated in the browser, a file containing customer records or an internal export can be converted without being uploaded to a third party.

Use the indentation control to suit the destination: two spaces is easy to read, four matches the PHP, Python and Java defaults, and a tab suits an editor set up that way. Indentation is purely cosmetic and never changes the meaning of the document.