Turn CSV into JSON without losing the details

Prepare a small, inspectable source

Use a UTF-8 CSV file under 2 MB with at most 10,000 data rows. The first row must name every column, and each name must be unique. Each following row needs the same number of fields.

Try our sample CSV, which includes an identifier with a leading zero and a title containing a comma:

name,region,code
Field notes,Europe,001
"A new, good idea",Asia,002
Next step,Americas,003

The quotation marks around the second name keep its comma inside one field. If that name were not quoted, the row would contain four fields instead of three.

Decide what the values mean

The CSV to JSON tool keeps every value as a string. That is deliberate: a value such as 001 may be an identifier, not the number 1. A date such as 03/04/2026 has different meanings in different locales. Guessing would hide an important decision.

An empty field becomes "", not null. A value true remains "true", not a boolean. If your destination requires numbers or booleans, convert the relevant columns explicitly in that system after reviewing the schema.

If your spreadsheet already changed 001 to 1, exporting to JSON cannot restore the missing zero. Check the source cells and export identifier columns as text.

Convert the file

1. Open CSV to JSON.
2. Choose your file or paste CSV into CSV content.
3. Choose its delimiter. Comma, semicolon, tab and pipe are supported. The tool does not guess the delimiter.
4. Select Convert to JSON.
5. Read the row and column count. Inspect the output preview.
6. Choose Download result to save the complete JSON array.

The quoted name and identifier above become:

{
  "name": "A new, good idea",
  "region": "Asia",
  "code": "002"
}

Headers have leading and trailing whitespace removed. Value whitespace is preserved. Blank physical lines outside quoted fields are skipped; a row containing empty fields separated by delimiters is retained.

Handle quotes and line breaks correctly

Inside a quoted field, write a literal quotation mark twice. For example, "Say ""hello""" becomes Say "hello". A line break inside a properly quoted field remains part of that value. Both common Windows and Unix line endings are supported.

After the closing quote, place the delimiter or line ending directly next to it. The parser rejects stray text and spaces there rather than silently changing the input. If a quotation mark appears inside an unquoted field, quote the entire field and escape the inner quotation mark.

Verify before importing

Compare three rows with the source: the first, one with unusual characters and the last. Check column names, non-Latin characters, identifiers, empty values and multiline text. Confirm that the destination accepts an array of objects rather than one object per line.

The preview may be shortened for long output, but the downloaded file contains all converted rows. This tool does not validate a destination’s schema or perform a data import.

If conversion fails, start with Troubleshoot CSV imports. Files and pasted text stay in this browser tab; closing it discards the workspace.