CSV to JSON Converter
And JSON Back to CSV

Paste the CSV into the input box, drop a .csv or .tsv file onto it, or click Open File. The JSON appears instantly on the right. By default the first row becomes the property names and every other row becomes one object in an array. Copy the result or download it as a .json file.

Yes. The parser follows RFC 4180: a value wrapped in double quotes can contain the delimiter, line breaks, and quotes written as two double quotes. If a quote is never closed, the converter reports the line where the quoted value started instead of silently merging rows.

Yes. Auto-detect tries comma, tab, semicolon, and pipe, ignoring any that appear inside quoted values, and picks the one that splits the first rows into the most consistent number of columns. The status line shows which delimiter was used, and you can override it or enter any single custom character.

Values that would change if turned into numbers are kept as text. A ZIP code like 00501 would lose its leading zeros, and an ID like 9007199254740993 is larger than JavaScript can store exactly, so both stay strings. Turn off type detection to keep every value as a string.

Name the columns with dots, such as address.city and address.zip, and turn on "Nest dot.notation headers". Each row then becomes an object with a nested address object. Going the other way, "Flatten nested objects" in JSON to CSV mode writes the same dot.notation columns, so the two directions round-trip.

By default they are skipped and listed in a report with the row number, source line, expected and actual value counts, and a preview, so nothing is silently shifted into the wrong column. If you would rather keep them, choose "Keep them" in Options: missing values become null and extra values get column_N keys.

Yes. Switch to JSON to CSV and paste an array of objects, a single object, or newline-delimited JSON. Turn on Windows line endings for the most compatible file, and "Neutralize spreadsheet formulas" when the data came from users, so a value starting with = cannot run as a formula when the file is opened.

By default a nested object or array is written into its cell as JSON text, the same as the JSON Tools converter on this site, so no data is lost. With flattening on, nested objects become dot.notation columns while arrays stay JSON text. When read back with type detection on, that JSON text becomes a real array or object again.

There is no fixed limit - the only limit is your device's memory. Files of several megabytes convert in a moment. For very large inputs, live conversion pauses automatically and you click Convert when ready, and the preview table shows the first 100 rows so the page stays responsive.

No. Files are read by your browser and converted with JavaScript on your device. Nothing is uploaded, logged, or stored, so it is safe to use with customer lists, financial exports, and other confidential data.