🧾 JSON to CSV Converter

How the Conversion Works

At its core, the converter runs a single JSON.parse() call on your input and then walks the resulting structure to build rows. Before anything is written, the tool verifies that the parsed value is a JSON array of objects. If you paste a bare object ({"a":1}), a string, a number, or an array of arrays, it stops and asks you for an array of objects instead — because only that shape has clear keys you can turn into a header row.

Once the input is validated, the header line is assembled from the keys of the first object, in exactly the order they appear in that object. Each subsequent row is built by looking up the same keys in every object, so the columns stay aligned across the whole table even when different objects define keys in different orders or omit some of them.

Every value then becomes one cell. Numbers, booleans, and plain text are written as-is. Anything that contains a comma, a double quote, or a line break is wrapped in double quotes, and any internal double quote is doubled ("") so it cannot be confused with the quoting itself — this follows the RFC 4180 escaping rules that Excel, Google Sheets, and most CSV importers expect. A tiny worked example makes this concrete:

Input :  [{"name":"Smith, John","city":"Amsterdam","score":9}]
Output:  name,city,score
         "Smith, John",Amsterdam,9

Note how Smith, John is quoted because of its comma, while Amsterdam and 9 are left untouched. That one rule — quote anything that could break the column structure — is what keeps the output valid for any downstream tool.

Worked Examples

Here are two examples with real, hand-verified data so you can see exactly what to expect.

Example 1 — simple records with a quoted field. Paste this order export:

[{"order_id":1024,"customer":"J. Smith","city":"Rotterdam","total":49.90}]

The first object defines four keys in this order — order_id, customer, city, total — so those become the header. The J. Smith value contains a comma, so that cell is quoted. The output is:

order_id,customer,city,total
1024,"J. Smith",Rotterdam,49.90

Example 2 — uneven keys with a missing field. Objects in a real dataset are not always identical. Consider this list of engineers:

[{"id":1,"name":"Ada"},{"id":2,"name":"Grace","role":"pioneer"}]

The columns follow the first object: id, name, role. The first row has no role, so that cell is simply left empty. The result stays aligned:

id,name,role
1,Ada,
2,Grace,pioneer

This is the key behaviour that keeps your table intact when later records carry extra fields — nothing shifts columns, and absent values become clean empty cells rather than errors.

Edge Cases & Limitations

A handful of situations are worth knowing about before you convert. Nested objects or arrays (a value that is itself a JSON object, such as an address or a tags list) are not flattened into extra columns — instead they are stringified into a compact JSON string inside the cell. No data is lost; the full nested value is visible in the spreadsheet, and you can split it out later if you need separate columns. null and undefined values become empty cells, matching the way empty fields behave in CSV.

Cells with a newline or a quote are quoted correctly, including doubling embedded quotes, so a multi-line address or a value like He said "hi" survives the round-trip intact. Large arrays — tens of thousands of rows — are processed locally in your browser with no upload and no server-side cap, so the only practical limit is your device's memory. UTF-8 characters such as café, Straße, or 東京 are preserved exactly, because the conversion works on JavaScript strings rather than mangling bytes.

About the Author

This converter is built and maintained by Aren Breur, the creator of ConvertAny.app. Aren specializes in privacy-first web tools that process your data entirely in the browser, so nothing is ever uploaded. Questions or feedback? Reach out at aren2connect@gmail.com.