Convert Text to CSV

Yes – for real, this time. Fields containing the output delimiter, a double quote, a carriage return, or a line feed get wrapped in double quotes; embedded quotes are doubled (say "hi" → "say ""hi"""). Many “text to CSV” tools skip this and produce broken CSV that breaks when a value has a comma in it. We don’t – the output passes through standard RFC 4180 parsers (Excel, Python csv module, Node csv-parse) without data corruption.

It scans the first non-blank line for delimiter characters in priority order: tab → comma → semicolon → pipe → multi-space (2+ consecutive spaces) → single space. Whichever appears first determines the input format. If your data is mixed (tabs on some lines, commas on others), auto-detect picks based on the first line – switch the drop-down to a specific format if that’s wrong.

Multi-space splits only on runs of 2 or more spaces – useful when your data has fields like "New York" that contain a single space. John 30 New York (double spaces) becomes 3 columns: John / 30 / New York. Single-space splits on any space, so New York becomes 2 columns – usually not what you want unless your values are guaranteed space-free.

Yes – click Swap. The reverse parser is a proper state machine: tracks “are we inside a quoted field?”, recognizes "" as an escaped quote inside a quoted field, allows commas and newlines inside quoted fields, and reports “unclosed quoted field” if the quote count is unbalanced. "Doe, Jr.",45,Boston parses to 3 fields, not 4.

Set input format to “Semicolon” and the parser splits on semicolons instead of commas – your decimal-comma values like 3,14 stay intact inside their fields. For output, pick “Semicolon (European CSV)” to round-trip cleanly to Excel in DE/FR/IT locales. The escaping logic uses the chosen output delimiter to decide what needs quoting.

Toggle it on and the table preview shows the first row as <thead> with styled header cells. The CSV output is unchanged – the header is still the first data row (that’s the RFC 4180 standard – headers aren’t a separate construct in CSV, they’re just the first row). The toggle is preview-only and survives Swap so reverse parsing also treats the first row as header in the preview.

200,000 characters. Beyond that the table preview slows browsers (each cell is a DOM node). The table itself caps at 50 displayed rows with a “Showing first 50 of N rows” note, but conversion processes the full input – only the preview is truncated. Copy and Download give you the complete output.

It doesn’t strip BOM specifically – BOM appears as a leading invisible character in the first field. If your CSV starts with  (BOM) followed by the actual data, the first parsed field will contain the BOM. Strip it from your source before pasting if needed, or untick “Trim whitespace” to see it exactly. Most modern CSV exports don’t include BOM unless explicitly requested for Excel compatibility.

No. All parsing, escaping, RFC 4180 state machine, and TSV/CSV 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, or anything you’d rather not send to a third-party converter.

Yes. Total bundle is about 22 KB. Load once, disconnect, keep converting. The parser and serializer are pure JavaScript – no library dependencies, no fonts to fetch, no analytics. Useful for cleaning data on airgapped systems, planes, or anywhere a cloud-based converter wouldn’t reach.