CSV to TSV is a our table engine conversion that rewrites comma-delimited records as tab-delimited text while still parsing CSV quotes before the delimiter changes.
From real runs
| Element | Result | What happens |
|---|---|---|
| Comma delimiters | Changed | Field boundaries move from commas to tab characters after parsing, so commas inside quoted CSV cells stay as cell text. |
| Quoted multiline cells | Kept | A CSV note containing an actual line break stayed one TSV field and was quoted in the output to protect the embedded newline. |
| UTF-8 BOM | Changed | The leading byte order mark was removed before the first header was written, leaving id instead of an invisible-prefixed name. |
| Numeric-looking IDs | Changed | Values such as 001 and 02134 were emitted as 1 and 2134 because scalar inference currently treats them as numbers. |
| ISO dates | Changed | A 2026-09-10 cell became 2026-09-10T00:00:00.000Z, which is clearer for machines but different from the source text. |
| Ragged rows | Changed | A short row was padded with empty TSV fields, and the API returned a ragged-row warning for the missing cells. |
Field boundaries move from commas to tab characters after parsing, so commas inside quoted CSV cells stay as cell text.
A CSV note containing an actual line break stayed one TSV field and was quoted in the output to protect the embedded newline.
The leading byte order mark was removed before the first header was written, leaving id instead of an invisible-prefixed name.
Values such as 001 and 02134 were emitted as 1 and 2134 because scalar inference currently treats them as numbers.
A 2026-09-10 cell became 2026-09-10T00:00:00.000Z, which is clearer for machines but different from the source text.
A short row was padded with empty TSV fields, and the API returned a ragged-row warning for the missing cells.
We converted our CSV sample (226 B) with Table engine on 2026-09-11. It took 1 ms and produced a 281 B TSV file, 1.2× larger than the input. The engine raised no warnings.
Input · sample.csv · 226 B
4 data rows × 5 columns; header: company,stage,arr,owner,close_date
Output · sample.tsv · 281 B
4 data rows × 5 columns; header: company stage arr owner close_date
First lines of the output
company stage arr owner close_date
Northwind Traders Proposal 48000 R. Ayers 2026-10-14T00:00:00.000Z
Contoso Ltd Negotiation 120000 P. Ng 2026-09-30T00:00:00.000Z
Fabrikam Qualified 22500 R. Ayers 2026-11-02T00:00:00.000Z
Tailspin Toys Won 64000 M. Okafor 2026-09-03T00:00:00.000ZPython csv module
python - <<'PY'
import csv
with open("input.csv", newline="", encoding="utf-8-sig") as src, open("output.tsv", "w", newline="", encoding="utf-8") as dst:
reader = csv.reader(src)
writer = csv.writer(dst, delimiter="\t")
writer.writerows(reader)
PYThis preserves text literally; it will not reproduce tabular-ir's number, boolean, null, or date coercion.
The HTTP API and the MCP server run the exact engine this page uses (Table engine); the response carries the same warnings the web converter shows. Check support first with GET /api/v1/pairs?from=csv&to=tsv.
curl -X POST https://filetypeconverters.com/api/v1/convert \
-H "Authorization: Bearer $FTC_API_KEY" \
-F "file=@input.csv" -F "from=csv" -F "to=tsv" \
-F 'options={"delimiter":","}' \
--output output.tsv| Key | Applies to | Type | Default |
|---|---|---|---|
| delimiter | Reading CSV | string | , |
| header | Reading CSV | boolean | true |
| quote | Reading CSV | string | " |
| encoding | Reading CSV | utf-8 · latin1 | utf-8 |
| delimiter | Writing TSV | string | , |
| newline | Writing TSV | · |
No. The CSV parser reads quoted cells before TSV is written, so Ada, Inc. stayed one field in our run. A quoted line break also stayed inside the note cell instead of creating an extra record.
The current tabular engine infers simple numeric strings as numbers. That helps totals and counts, but it is unsafe for ZIP codes, invoice IDs, and codes with leading zeroes.
The route resolver marks csv-to-tsv as a table engine browser pair, so the web converter can process the file on the device. API requests still upload the file because the API endpoint itself is server-side. The conversion counts toward your daily or monthly allowance like every other, and the input is deleted within 24 hours.