JSON to JSONL emits one compact object per line, but our table engine flattens nested JSON before writing the newline-delimited records.
From real runs
| Element | Result | What happens |
|---|---|---|
| Top-level array | Changed | The surrounding JSON array brackets were removed and each record was written on its own line. |
| Record count | Kept | Two input JSON objects produced exactly two JSONL lines in the downloaded file. |
| Nested objects | Changed | customer.name, customer.city, and extra.gift became flat keys in each line object. |
| Arrays | Changed | The tags arrays became JSON-encoded strings inside each line object rather than native arrays. |
| Null padding | Changed | The first line received extra.gift:null because another record introduced that column. |
| String and Unicode values | Kept | String IDs with leading zeros and non-ASCII city names were preserved in the JSONL output. |
The surrounding JSON array brackets were removed and each record was written on its own line.
Two input JSON objects produced exactly two JSONL lines in the downloaded file.
customer.name, customer.city, and extra.gift became flat keys in each line object.
The tags arrays became JSON-encoded strings inside each line object rather than native arrays.
The first line received extra.gift:null because another record introduced that column.
String IDs with leading zeros and non-ASCII city names were preserved in the JSONL output.
We converted our JSON sample (557 B) with Table engine on 2026-09-11. It took 1 ms and produced a 411 B JSON Lines file, 26% smaller than the input. The engine raised no warnings.
Input · sample.json · 557 B
Array of 4 objects; keys: company, stage, arr, owner, close_date
Output · sample.jsonl · 411 B
4 JSON objects, one per line; keys: company, stage, arr, owner, close_date
First lines of the output
{"company":"Northwind Traders","stage":"Proposal","arr":48000,"owner":"R. Ayers","close_date":"2026-10-14"}
{"company":"Contoso Ltd","stage":"Negotiation","arr":120000,"owner":"P. Ng","close_date":"2026-09-30"}
{"company":"Fabrikam","stage":"Qualified","arr":22500,"owner":"R. Ayers","close_date":"2026-11-02"}
{"company":"Tailspin Toys","stage":"Won","arr":64000,"owner":"M. Okafor","close_date":"2026-09-03"}Python json module
python - <<'PY'
import json
with open("input.json", encoding="utf-8") as f:
rows = json.load(f)
with open("output.jsonl", "w", encoding="utf-8") as out:
for row in rows:
out.write(json.dumps(row, ensure_ascii=False, separators=(",", ":")) + "\n")
PYThis direct dumper keeps nested arrays and objects; tabular-ir flattens them first.
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=json&to=jsonl.
curl -X POST https://filetypeconverters.com/api/v1/convert \
-H "Authorization: Bearer $FTC_API_KEY" \
-F "file=@input.json" -F "from=json" -F "to=jsonl" \
-F 'options={"encoding":"utf-8"}' \
--output output.jsonl| Key | Applies to | Type | Default |
|---|---|---|---|
| encoding | Reading JSON | utf-8 · latin1 | utf-8 |
| flattenSeparator | Reading JSON | string | . |
| arrayMode | Reading JSON | json · explode | json |
| delimiter | Writing JSON Lines | string | , |
| newline | Writing JSON Lines | · | |
| sheetName | Writing JSON Lines | string | |
| dialect | Writing JSON Lines | postgres · mysql · sqlite | postgres |
| tableName | Writing JSON Lines | string | data |
| createTable | Writing JSON Lines | boolean | true |
No. It changes the container from one array document into one JSON object per line. The output is meant for stream processors and log pipelines.
Not as arrays in this route. The default tabular conversion serializes arrays to strings before writing JSONL lines.
Yes. The parser treats a top-level object as one record, so the output becomes one JSONL line after flattening.