CSV ↔ JSON Converter (RFC 4180 compliant)
Paste CSV or JSON. The converter parses RFC 4180-compliant CSV (handles quoted fields with commas, escaped quotes, multi-line values). For JSON-to-CSV, accepts both array-of-objects (uses keys as headers) and array-of-arrays.
[
{
"name": "Alice",
"age": "30",
"city": "Tokyo"
},
{
"name": "Bob",
"age": "25",
"city": "Seoul"
},
{
"name": "Carol",
"age": "42",
"city": "Madrid"
}
]How it works
Why RFC 4180 matters
CSV looks simple but has edge cases: a field containing a comma must be quoted; a field containing a quote must escape it as a doubled quote; a field can contain newlines if quoted. Naive split-on-comma parsers break on real-world data. This converter implements the RFC 4180 grammar exactly.
Naive parsers also break on Excel exports that use semicolons (common in European locales where ',' is decimal separator) or tabs (TSV). The delimiter dropdown handles all four common variants.
JSON shapes for output
Array of objects (default with header row): each row becomes {column: value}. The most common JSON shape, easy to consume in any language.
Array of arrays (header row off): each row becomes [v1, v2, ...]. Useful when columns aren't named or you want a positional structure.
Going JSON to CSV, the converter detects which shape you have. Objects: extracts unique keys across all rows for headers. Arrays: writes rows as-is.
Common pitfalls
Excel locale: in some European Excel installs, the default delimiter is ';' not ','. If your CSV looks like one giant column when imported, switch to ; in the delimiter dropdown.
BOM (Byte Order Mark): some Excel exports prefix the file with U+FEFF. We pass it through; if your downstream parser fails, strip the first 3 bytes.
Trailing newline: a CSV ending in '\n' adds an empty last row in some parsers; we filter empty rows automatically.
Numbers as strings: CSV has no types — '42' becomes the string "42" in JSON. If you need typed values, run a post-processing step that coerces known number columns.
Frequently asked questions
›Does this handle quoted fields with commas?
Yes — RFC 4180 quoting is fully supported. "Doe, John" is parsed as a single field.
›How are escaped quotes handled?
Per RFC 4180, a literal '"' inside a quoted field is written as '""' (two quotes). The parser handles this correctly.
›Can fields contain newlines?
Yes, when quoted. "line 1\nline 2" inside quotes is treated as one field with embedded newline.
›Does this support BOM (UTF-8)?
BOM passes through. If your downstream consumer doesn't tolerate it, strip the first 3 bytes (EF BB BF) from the CSV.
›What about huge files?
Browser handles a few MB without issue. For very large CSVs (100MB+), use a streaming parser like Papa Parse with a worker, or a CLI tool.
›Are numbers preserved as numbers?
CSV → JSON: all values become strings (CSV is untyped). JSON → CSV: numbers become unquoted CSV cells. To get typed JSON, post-process columns you know are numeric.
›Why does my JSON show empty strings instead of null?
CSV represents missing values as empty cells. We map them to empty strings. To get null in JSON, post-process the output.
›Does the data leave my browser?
No. Conversion runs locally; nothing is sent to a server.
Related tools
Last updated: