How to open JSONL and NDJSON files

JSON array versus line-delimited records

The same events can be written as a normal JSON array or as JSONL/NDJSON. In the array version, commas and brackets define one top-level value. In the line-delimited version, each non-empty line is its own JSON value.

The names are usually practical aliases in developer workflows: .jsonl is often called JSON Lines, while .ndjson expands to newline-delimited JSON. JSON All-in-One recognizes .jsonl and .ndjson as strict line-delimited suffixes.

[
  { "id": "evt_1", "type": "signup", "user": { "id": "u_1" } },
  { "id": "evt_2", "type": "paid", "user": { "id": "u_1" } }
]

{"id":"evt_1","type":"signup","user":{"id":"u_1"}}
{"id":"evt_2","type":"paid","user":{"id":"u_1"}}

• Open larger record exports

Valid records and common mistakes

A newline inside a JSON string must be escaped as \n; it is not a physical line break inside the record. A blank line can be ignored by the normalizer, but a partial object line should fail because it is not a complete JSON value.

JSON All-in-One also has a relaxed NDJSONC mode for .jsonlc and .ndjsonc, where JSONC comments and trailing container commas are accepted. Treat that as a separate relaxed format. If a data partner asked for strict NDJSON, send strict records.

{"id":"evt_3","message":"first line\nsecond line"}

{"id":"evt_4","message":"first line
second line"}

Open the file and move between records

Open a .jsonl or .ndjson file from Chrome or from the JSON All-in-One workspace. With no content-type hint, the detector needs at least two complete records before a line-delimited sniff is conclusive. A source that starts with [ is treated as an ordinary JSON or JSONC array, not as NDJSON.

Once loaded, the workspace model is an array. The first record is $[0], the second is $[1], and nested fields are selected from those array items. That is why a JSONPath like $[?@.type == "paid"].user.id makes sense in the product.

$[?@.type == "paid"].user.id

map(select(.type == "paid") | {event: .id, user: .user.id})

• Filter normalized records with JQ

When conversion helps

JSONL can be read in the workspace without first converting it to a pretty JSON array. Conversion is useful when another tool expects one JSON document, when you want to generate a schema from the combined samples, or when an export needs a root array.

The reverse is also true: export back to line-delimited records only when the receiving system expects one record per line. That export workflow is separate from opening and reading JSONL.

• Use JSONPath for selection when you only need matched values.
• Use JQ when you need to reshape records before export.
• Use schema generation when several records show the optional and nullable fields you care about.

Troubleshooting

If only one line is present, the input may be valid JSON but not a useful line-delimited sample for sniffing. Add a second complete record or use a recognized .jsonl or .ndjson suffix. If a record reports a line number error, inspect that line as a complete JSON value first.

If a copied query fails, confirm that it targets an array. Many CLI JQ examples for NDJSON use streaming or slurp flags outside the expression itself. In JSON All-in-One, write the expression for the normalized array model.

Next step

After the file opens, the most useful next action is usually a small filter: find the event type, customer id, status, or error code that started the investigation. For transformation examples, continue with JQ examples for filtering and transforming JSON.