The result is not valid JSON as a whole
There is no enclosing array and no commas between records, so a standard JSON parser will reject the file. It must be read line by line — which is the entire point of the format.
Convert CSV to JSONL
JSON Lines is what bulk loaders, log pipelines and training sets expect: one self-contained JSON object per line, streamable without loading the whole file. This converter turns a CSV into exactly that, with real types rather than strings everywhere.
- Runs in your browser
- No signup
- No upload
- Free
CSVJSONL
Runs on your deviceCSV
Options
Reading CSV
Auto-detect reads the first rows and picks the separator that splits them consistently.
Turn off when the file starts straight into data. Columns are then named column_1, column_2, and so on.
Turns "42" into a number and "true" into a boolean. Values with leading zeros stay text so IDs and zip codes survive.
Which worksheet to convert. Filled in once the workbook is read.
Dotted path to the array of records, such as data.items. Leave empty to use the whole document.
Turns {"user":{"id":7}} into a user.id column. Arrays are kept as JSON text.
Pick the encoding the file was written in. Wrong encoding shows up as garbled accents.
Comma-separated words that should become null instead of text.
Writing JSONL
Snappy is the default across Spark, DuckDB and pandas. Gzip is smaller but slower to read.
Turns a user.id column back into {"user":{"id":7}}.
Excel needs a byte order mark to open UTF-8 files with accents correctly.
Rows per row group. Larger groups compress better; smaller groups let readers skip more.
Converting…
Ad slot ·
conversion-preroll
This short ad is what keeps the converter free.
Work with Parquet as well?See ParquetReader with ParquetReader — opens Parquet files directly.What happens
How this converter handles CSV to JSONL
- Parses the CSV with automatic delimiter detection and full quoted-field handling.
- Converts values to real JSON types — numbers, booleans and nulls, with leading zeros kept as text.
- Writes one compact JSON object per line, with no wrapping array.
- Can rebuild nesting from dotted headers so each line carries a nested object.
Format differences
CSV vs JSONL
Plain text rows that open anywhere, with no types and no schema. One JSON object per line, built for streaming and appending.
| CSV | JSONL | |
|---|---|---|
| Type information | None — every value is text | Strings, numbers, booleans, null |
| Structure | Flat rows and columns | One nested object per line |
| Compression | None | None |
| Typical file size | Baseline (100%) | Slightly under JSON |
| Schema | Header row at best | Implicit, per line |
| Human-readable | Yes, in any text editor | Yes, one record at a time |
| Tool support | Universal | Log pipelines, LLM training sets, BigQuery |
Worth knowing
What to watch out for
Every line repeats the keys
JSONL is larger than the CSV it came from, because each record carries its own field names. That is the cost of every line being independently parseable.
Empty fields become null, not empty string
An empty CSV cell has no way to distinguish "blank" from "missing". It becomes null here. Turn off type detection if you need empty strings preserved as strings.
When you need this
Common reasons to convert CSV to JSONL
Bulk loading into BigQuery
Newline-delimited JSON is a native BigQuery import format.
Preparing fine-tuning or evaluation datasets
JSONL is the standard format for model training data.
Feeding a streaming consumer
A line-based reader can start work immediately rather than waiting for the full file.
FAQ
CSV to JSONL questions
Is my CSV uploaded?
No. The conversion runs in your browser using WebAssembly and JavaScript. The file is read from your disk into a Web Worker, converted in memory, and handed back as a download. It is never sent to a server, so there is nothing for us to store, log or leak.
What is the difference between this and CSV to JSON?
JSON produces one array containing every record, which a parser must read entirely. JSONL produces one object per line, which can be streamed and appended to. Pick JSONL for pipelines and JSON for documents.
Is the output minified?
Yes. Each line is a compact JSON object, since pretty-printing would break the one-record-per-line contract.
Can each line be a nested object?
Yes — use dotted headers and enable "Rebuild nested objects".
Which extension should I use?
Either .jsonl or .ndjson. They are interchangeable; .jsonl is the more common choice today.
Keep going