CSV and JSON Converter

How to convert CSV to JSON

1. Paste CSV or import one UTF-8 CSV/TSV file.
2. Select comma, semicolon or tab and choose whether the first record contains column names.
3. Press Convert to JSON, check the record preview, then copy or download the complete array.

A string-preserving example

name,code
Alice,00123

The resulting JSON keeps the identifier's leading zeros:

[
  {
    "name": "Alice",
    "code": "00123"
  }
]

CSV to JSON rules

Cells are not trimmed or interpreted as numbers, booleans, null or dates. Headers are trimmed and must be non-empty and unique. Empty cells are empty strings. Header-only input produces []. A single initial UTF-8 byte order mark is removed during parsing.

Quoted commas, escaped quotes and embedded newlines are supported. Invalid quoting and inconsistent field counts produce a parser-record error while preserving the source. CSV records are not necessarily physical lines: do not deduplicate arbitrary CSV with a raw line-removal tool.

CSV to JSON is limited to 5 MiB UTF-8 and 50,000 data rows, 4,096 columns and 32 MiB output; no data is silently truncated. It does not infer cell types or build nested mappings.

How to convert JSON to CSV

1. Choose JSON to CSV. Paste an array of flat objects or import one UTF-8 .json file.
2. Choose a delimiter, optional UTF-8 BOM and formula protection. Press Validate / Convert to CSV.
3. Select or remove detected columns and reorder them with the keyboard-accessible Move up/down buttons. Convert again after any change.
4. Review the first 50 rows and protection changes, then copy or download the complete converted.csv or converted.tsv file.

[{"name":"Alice","code":"00123"},{"name":"Bob","active":true}]

Columns follow first appearance across objects. This example exports:

name,code,active
Alice,00123,
Bob,,true

JSON to CSV rules and limitations

Only an array of flat objects is accepted. Nested objects/arrays, null or primitive rows, duplicate keys, invalid JSON and unpaired Unicode surrogates produce actionable errors. Empty arrays and arrays of empty objects validate, but export is disabled until columns exist. Numeric-looking keys retain encounter order; prototype-like names are handled safely.

Numeric tokens are preserved exactly, including 9007199254740993, long decimals, negative numbers and exponent spellings. Strings keep their content unless formula protection adds an apostrophe. Booleans serialize as lowercase true/false. Null, missing keys and empty strings become empty CSV fields, losing their distinction.

Formula protection is on by default. It prefixes an apostrophe to potentially formula-like strings and headers; it does not modify numeric or boolean values. Review the count and change preview. Disabling protection requests raw text explicitly. Quoting alone does not neutralize spreadsheet formulas, and no CSV is universally safe for every spreadsheet application.

CSV quoting escapes delimiters, quotes and embedded line breaks; record separators are CRLF. A header row is always included. Export delimiters can be comma, semicolon or tab. A UTF-8 BOM is optional and off by default; spreadsheet encoding detection varies.

JSON to CSV accepts 5 MiB input, 50,000 records and 200 distinct columns, with a 20 MiB output safeguard. The table previews 50 rows and the text area previews 8,192 characters; copy and download include the full accepted export. Clipboard denial exposes selectable complete output. Changing settings or columns invalidates the old export.

This is not a lossless JSON/CSV round trip: types and null distinctions are lost, CSV-to-JSON headers are trimmed, protection changes strings, and spreadsheet applications can reinterpret leading zeros or dates. Neither direction reads Excel workbooks, flattens nested JSON, supports JSON Lines or uploads data. Cancel, clear and undo keep control of in-memory drafts.

Frequently asked questions