Data Format Converter

CSV · TSV · JSON · NDJSON converter

Paste your data, pick a target format, and convert it — CSV, TSV, JSON, and NDJSON (JSON Lines) in any of the twelve directions. Nothing is uploaded: the conversion is compiled to WebAssembly and runs locally in your browser tab, so your data never leaves your device.

How it works

• From — leave it on Auto-detect and the tool sniffs the input: text starting with [ is a JSON array, { is a single JSON object (or NDJSON when every line is its own JSON value), and anything else is parsed as CSV — or TSV when the first row is tab-separated. Force a source format if you need to.
• To — choose JSON (a single array), NDJSON / JSONL (one compact record per line, ideal for streaming, log pipelines, MongoDB mongoimport, BigQuery, and JSONL training sets), CSV, or TSV.
• First row is headers — for CSV/TSV, treat the first row as field names. On (the default) you get an array of objects keyed by the headers; off gives an array of arrays. When writing CSV/TSV it emits a header row from the union of every record's keys, in first-seen order, filling blanks where a row is missing a key.
• Infer types — CSV/TSV cells become JSON numbers, true/false, and null for empty cells. Turn it off to keep every cell as a string. Values that would not round-trip — leading zeros like 007, a leading + — always stay strings, so zip codes and phone numbers survive intact.
• Pretty-print — indent the JSON target. NDJSON is always one compact record per line, so this option does not affect it.
• Flatten — when writing CSV/TSV, expand nested objects and arrays into dot-notation columns ({"addr":{"city":"NYC"}} becomes an addr.city column) instead of writing them as compact JSON strings.

Quoted fields, embedded commas/tabs, and newlines are handled per RFC 4180.

Worked example

Input CSV:

name,age,active
Alice,30,true
Bob,25,false

To NDJSON →

{"name":"Alice","age":30,"active":true}
{"name":"Bob","age":25,"active":false}

To JSON → [{"name":"Alice","age":30,"active":true},{"name":"Bob","age":25,"active":false}]

Going the other way, {"a":1,"b":2} / {"a":3,"c":4} as NDJSON → CSV gives a,b,c with a blank cell where the second row has no b.

Limits & scope

• Everything runs in your browser, so the practical size limit is your device's memory — there is no fixed cap, but multi-hundred-MB files may be slow.
• This tool covers the four named formats. For semicolon- or pipe-delimited files, or arbitrary delimiter swaps, use a dedicated delimiter converter.
• Flattening applies to CSV/TSV output; it does not rebuild nested objects from dot-notation headers on input.

FAQ