CSV ↔ JSON Converter

The parser follows RFC 4180: fields are separated by the delimiter, a field in double quotes may contain the delimiter, line breaks and quotes doubled as "", and CRLF or LF end a record. The delimiter is detected as the most frequent of comma, semicolon, tab and pipe in the first ten lines — choose it explicitly when values contain that character unquoted. A stray quote or an unterminated quoted field is reported with its line rather than swallowed. A leading byte-order mark is removed.

CSV has no types; everything is text. With typing on, a cell becomes a number only when it is a strict decimal literal without leading zeros and within 2⁵³ (so "007", "00123" and long IDs stay strings), true/false become booleans, and an empty cell becomes null. The page counts how many cells were converted. Turn typing off to keep every value as a string — the safe choice for postcodes, phone numbers and anything with a leading zero.

An array of objects becomes a header of every key in first-seen order and one row per object, with missing keys as empty cells; an array of arrays becomes positional rows. A nested object or array cannot be a flat cell, so it is written as JSON text inside the cell and the page says how many. Fields with the delimiter, quotes, line breaks or surrounding spaces are quoted per RFC 4180.

Spreadsheet exports vary: Excel may write a BOM and use the locale's list separator (semicolon where the decimal comma is used), and dates arrive as whatever text the sheet produced — the page keeps them as text. Encoding is assumed UTF-8. Files up to 5 MB are converted on your machine; nothing is uploaded. The same four anonymous usage counts as the rest of the site apply.