How to Convert JSON to JSONL (JSON Lines) — With Code Examples & Free Tool

JSONL (JSON Lines) puts one JSON object per line — no wrapping array, no commas between records. It's the standard format for LLM training data, log streaming, and big data pipelines.

JSON → JSONL

Python

import json

with open("data.json") as f:
    records = json.load(f)  # expects a JSON array

with open("data.jsonl", "w") as out:
    for record in records:
        out.write(json.dumps(record) + "\n")

jq

jq -c '.[]' data.json > data.jsonl

The -c flag outputs compact (single-line) JSON. .[] iterates array items.

JSONL → JSON (reverse)

Python

import json

records = []
with open("data.jsonl") as f:
    for line in f:
        records.append(json.loads(line.strip()))

with open("data.json", "w") as out:
    json.dump(records, out, indent=2)

When to use JSONL

Use case	Why JSONL wins
LLM training data	One record per line = easy streaming
Log files	Append without modifying existing JSON
Big data (Spark, BigQuery)	Line-by-line processing = no full load
API streaming	Send records as they're ready

No-code conversion

Use the free JSON to JSONL tool — paste or upload your JSON array and download the JSONL file instantly. The JSONL to JSON tool handles the reverse.