JSON to CSV Converter

API responses are rarely uniform. The first object may carry an id and a name, while a later one adds a status and a nested address. Taking the keys of the first object alone would silently drop everything the others have, so the tool collects the union of every key it meets, in the order the keys first appear.

Column order therefore reflects your document rather than alphabetical sorting, which usually makes the result easier to scan. A record missing a column gets an empty cell, and the number of such records is reported so an inconsistent payload does not pass unnoticed.

A nested object has no natural place in a spreadsheet, so with flattening on {"address":{"city":"Leeds","postcode":"LS1"}} produces address.city and address.postcode. The dotted path keeps the origin of the value visible, which matters once the sheet has columns from several nested objects.

Turn flattening off and the nested object is written as compact JSON inside a single cell instead. That is the better choice when the nested structure is itself meaningful, such as a list you want to open in a text editor later.

Flattening deliberately stops at lists. A list has no key to hang dotted names on, so expanding "tags":["a","b"] would invent columns such as tags.0 whose meaning depends on the order of the array. Instead the list is kept as compact JSON in one cell, which survives the round trip and can be split into columns afterwards if you actually need that.

Comma is the default, but a spreadsheet in a European locale will not split on it, so semicolon, pipe and tab are available. Pick the same delimiter the file will be opened with rather than the one that looks tidier in a text editor.

Values are written the way JSON spells them: true, false, null and numbers without quotes. That keeps the file honest as data, though a spreadsheet displaying true as TRUE is just its own rendering. Line endings default to LF, with CRLF available for older Windows tools that still expect it.

An array of objects is the common case and becomes one row per element. A single object becomes a single row. A plain list of values, such as [1,2,3], becomes a single value column, because there are no field names to invent and rows of data with no headers are still useful data.

Anything that holds no values, such as an empty array, is reported as having nothing to write rather than producing a file containing only a blank line.