Convert TSV to JSON

Yes – that’s the type inference. "42" becomes the JSON number 42, "3.14" becomes 3.14, "1e3" becomes 1000, "true"/"True"/"TRUE" all become boolean true, "null" becomes JSON null. The regex for numbers is strict – no commas allowed (locale ambiguity), no spaces, no multiple dots. Values like "1,000" or "$30" stay strings because they’re not unambiguous numbers.

Three options: (1) Empty string: empty cells become "" – the key is present, value is empty string. (2) null: empty cells become JSON null – clearer “no value” signal. (3) Omit: the key is removed from the object entirely. So {"name":"Alice","age":""} becomes {"name":"Alice","age":null} or {"name":"Alice"} depending on which you pick. Useful when receiving systems treat these cases differently.

Yes – turn off “First row is headers”. Output becomes an array of arrays where each row is one inner array: [["name","age"],["Alice",30],["Bob",25]]. Type inference still applies to individual values. Useful when your TSV genuinely has no header row, or when you want headers preserved as data rather than as keys.

TSV uses tab separators with no quoting convention, so the parser is simpler – just split each line on tab characters. CSV has RFC 4180 escaping rules for embedded commas/quotes/newlines, which requires a stateful parser. If your data has fields with embedded tabs, TSV won’t work – you’d need CSV. If your data has clean tab-separated fields, TSV is faster to parse and the input is more compact.

For an array of objects, headers are the union of all object keys in order of first appearance. So [{"a":1},{"a":2,"b":3}] produces headers a then b – keys that appear later still make it into the header row. For an array of arrays, no headers are inferred; rows go straight to output. Nested objects/arrays get JSON-stringified into a single TSV cell.

They stay strings. The number regex doesn’t accept commas because locale rules differ – US uses 1,000 for one thousand, Europe uses 1.000. Rather than guess, the tool preserves the original. If you want them as numbers, pre-process your TSV to use standard decimal points before pasting.

They’re JSON-stringified into a single TSV cell. So [{"name":"Alice","address":{"city":"NYC"}}] becomes nametaddressnAlicet{"city":"NYC"}. TSV can’t represent nested structure natively, but stringifying preserves the data so a round-trip through both directions is possible if you parse the address cell as JSON afterwards.

500,000 characters. That’s typically 10,000+ rows of moderate-width data. Beyond that, browsers slow down for the JSON.stringify step and the textarea rendering. Hit the cap and you get a specific error pointing to the limit.

No. All parsing, type inference, JSON serialization run in your browser. Open DevTools → Network and confirm zero requests fire – even when you Convert or Download. Safe for proprietary datasets, customer lists, configuration data, or anything you’d rather not send to a third-party converter.

Yes. Total bundle is about 20 KB. Load once, disconnect, keep using. Pure JavaScript – no library dependencies (no Papa Parse, no jq), no fonts, no analytics. Useful for converting data on airgapped systems or in environments where you can’t share data with external tools.