Overview
This workspace parses JSON into a format-neutral value, validates and lints it, then serializes the same data as CSV.
Convert JSON to CSV with live validation, linting, and browser-local processing.
This workspace parses JSON into a format-neutral value, validates and lints it, then serializes the same data as CSV.
JSON objects, arrays, strings, numbers, booleans, and null-like values are mapped to the closest compatible CSV representation.
Duplicate keys, unsupported top-level TOML values, unsafe integers, aliases, and format-specific date types can prevent exact cross-format equivalence.
Processing occurs in a browser worker. Input, output, keys, and values are never uploaded or included in analytics events.
Quoted fields, doubled-quote escapes, and embedded newlines inside quotes are parsed per RFC 4180. Delimiter (comma, semicolon, tab, or pipe) is auto-detected from the first 20 records, or set it explicitly in the options.
Numbers, true/false, and null are inferred from unquoted cells — but a leading zero ("007") and integers beyond JavaScript's safe range are deliberately kept as strings, since converting them would silently change the value.
Wrapping a cell in quotes (e.g. "42") opts it out of type inference entirely, even with inference enabled — useful for IDs or codes that happen to look numeric.
TOML has no top-level array type, so an array of CSV rows needs a wrapping table to convert to TOML — this shows as a clear diagnostic rather than a crash. Converting to JSON, YAML, or back to CSV needs no such wrapping.