JSON to CSV Converter

Example: a US order can store "currency": "USD" and "amount": 1234.56. JSON numbers use a dot and no thousands separator; write the date as "2026-10-11". Formatting preserves the stored value; it does not convert currencies or interpret local dates.

For exported data, record the units, date format and currency explicitly. A display value such as $1,234.56 needs a stated currency; 10/11/2026 needs a stated date convention. Check this tool’s output and required syntax before using it in another system.

Yes, completely free with no registration or file size throttling. All conversion happens locally in your browser.

An array of objects (each object becomes a row, keys become the header) and an array of arrays (each inner array becomes a row). A single object or any other top-level type is rejected with a clear message.

Objects do not need identical keys: the converter takes the union of all keys in order of first appearance, and missing values become empty cells. Nested objects and arrays are serialized as JSON strings in their cell rather than silently dropped.

Most likely a delimiter mismatch: Excel with European regional settings expects semicolons. Switch the delimiter to semicolon and convert again. If punctuation is garbled, that is an encoding issue on export - this tool outputs standard UTF-8.

No. The parser is plain JavaScript running on this page. Nothing is transmitted, logged, or stored.

Nested keys become dotted column names (user.address.city → user.address.city), and arrays of objects either expand to rows or to indexed keys (items[0].name) depending on the mode. The header is derived from the first record's shape, and fields missing in later records are left empty rather than erroring - the convention spreadsheet exports follow.

Strings containing commas, quotes, or newlines are wrapped and escaped per RFC 4180 automatically; null becomes an empty cell; booleans and numbers are written literally. Arrays and objects that could not flatten are serialized as JSON strings so no data is dropped.

Deeply nested or heterogeneous documents (every record a different shape) produce sparse CSVs with hundreds of columns. CSV suits tabular data with a consistent schema - if more than ~20% of cells would be empty, staying in JSON or loading into a proper analytical store is the better path.