Is my CSV uploaded?
No. Parsing happens in the browser tab. After the page loads it makes no network requests.
Drop a CSV, get clean JSON. Headers become keys, type inference included. Nothing uploaded.
Drop files here or
CSV looks simple until it is not. A field containing a comma needs quotes, a field containing a quote needs doubled quotes, a file saved by Excel on Windows starts with a byte-order mark that silently corrupts your first column name, and line endings may be CRLF. Naive splitters fail on all four, and they fail quietly: you get JSON that parses fine and is wrong.
This converter parses according to RFC 4180, strips the BOM, normalises CRLF, and infers numbers and booleans so the JSON is usable without a second pass.
The details that decide whether the output is usable downstream.
| Conversion behaviour | Behaviour | Notes |
|---|---|---|
| Delimiter | Comma, auto-detected alternatives | Semicolon and tab are detected automatically, which matters for European Excel exports. |
| Quoting | RFC 4180 | Commas, quotes, and newlines inside quoted fields are preserved correctly. |
| BOM | Stripped | Excel's UTF-8 BOM otherwise becomes part of your first key and breaks every lookup. |
| Line endings | CRLF and LF both accepted | Windows exports do not leave a trailing carriage return in the last column. |
| Type inference | Numbers and booleans | 123 becomes a number, true becomes a boolean. Header-only columns stay strings. |
| Output | Array of objects | One object per row, keyed by header. Empty cells become empty strings, not null. |
Excel re-quotes only what it thinks needs quoting and will happily mangle long numbers into scientific notation. Convert the original file.
A header like `First Name` becomes a key with a space. JSON allows it but most code does not want it. Rename headers in the CSV first.
Two columns called `id` produce one key: the second silently overwrites the first. Check the header row.
Flat CSV converts to flat JSON. Anything hierarchical has to be reshaped after conversion.
A CSV saved as Latin-1 with accented characters produces mojibake. Re-export as UTF-8 first.
No. Parsing happens in the browser tab. After the page loads it makes no network requests.
Comfortably into the tens of megabytes. The limit is your browser's memory, not this tool.
Yes, the delimiter is detected automatically, so European Excel exports work without changing settings.
They are preserved. The parser follows RFC 4180 rather than splitting on every comma.
Yes, numeric-looking values become JSON numbers rather than strings. Turn this off if you need IDs kept as strings.
Yes, there is a separate JSON to CSV tool on this site.
No, the byte-order mark is stripped before parsing.
Yes, download the JSON directly from the results panel.
FileKit Desktop runs every conversion locally as a single offline file — no 10-file limit, batch conversion, and no network connection required. One-time $19, no subscription.
See FileKit Desktop — $19