Check JSON Lines and NDJSON line by line, view records as a table, convert to a JSON array
Runs in your browser. Nothing you type is sent anywhere. Checked against sources on 16 Oct 2026.A JSONL file holds one JSON value per line. This viewer checks every line on its own, lists each bad line with its number, column and a fix, shows the records as a table with 1 column per key, and converts JSON Lines to a JSON array or back, keeping long numbers exact.
Open a .jsonl or .ndjson file, or paste the lines into the box.
Each line is checked as its own JSON value; a bad line is listed with its line number, column and a fix.
Read the records as a table, one column per key, or switch the output to a JSON array or clean JSON Lines.
Press Copy, or Download the converted file.
JSON Lines, also published as NDJSON, is the format of logs, analytics exports and machine-learning datasets: one JSON value per line, so a program can stream or append records without reading the whole file. An ordinary JSON viewer rejects such a file outright, because the file as a whole is not one JSON document.
What is checked. Each line is parsed on its own as RFC 8259 JSON. A bad line is listed with its line number, column and a likely fix, such as double quotes in place of single ones or a trailing comma removed, and the good lines still show. Windows line endings, a final newline and a byte order mark at the start are accepted, as the format allows.
Views.
| View | What you get |
|---|---|
| Table | The first 200 records, one column per key, nested values as JSON |
| JSON array | All records between brackets, each written exactly as in the file |
| Pretty JSON | The same array, indented |
| JSON Lines | Clean lines: trimmed, empty lines removed |
Converting both ways. Paste a JSON array instead and it is read as its elements, ready to save as JSON Lines for BigQuery, Elasticsearch or a fine-tuning upload. The JSON array view copies each record's own text, so a 20-digit ID is never rounded; when a view would round one, a note says so.
Empty lines. The NDJSON specification lets readers skip empty lines, so they are skipped and counted; a switch reports them as errors for strict readers.
Every line is judged as CPython's json module judges it, checked against 36 recorded edge cases. Nothing you open leaves the browser.
A JSONL file, also called JSON Lines or NDJSON, holds one complete JSON value on each line, usually one object per record. Logs, database exports and machine-learning datasets use it because a program can read or append one line at a time without loading the whole file.
Open the .jsonl or .ndjson file here, or paste its lines, and every record appears in a table with one column per key. Bad lines are listed with their line number and column. The file is read by your browser itself, so even a confidential export never leaves your device.
Switch the output to JSON array and the records are wrapped in square brackets with a comma after each, written exactly as they appear in the file. Pretty JSON indents them instead. Download saves the result as a .json file ready for any tool that expects one array.
Paste the JSON array, pretty or on one line, and it is read as its elements. Switch the output to JSON Lines and each element is written compactly on its own line, the layout that BigQuery, Elasticsearch bulk loads and fine-tuning uploads ask for.
The usual causes are single quotes instead of double quotes, a trailing comma before } or ], a value split over two lines, or NaN, which JSON does not allow. Each bad line is shown with its column and a suggested fix, while the valid lines still appear in the table.
Strictly, JSON Lines has a value on every line, but the NDJSON specification lets a reader skip empty lines, and most tools do. Empty lines are therefore skipped and counted, and a switch reports them as errors when a strict consumer will read the file.
Rates and rules change; check the source before you rely on a figure.
Still stuck? Ask about JSONL Viewer and Validator and anybody who knows can answer. No account needed.
Same job, other tools: JSON Validator – Check JSON and see exactly where it breaks · JSON Minifier – Strip whitespace from JSON for production · XML Validator – Check XML is well formed and see exactly where it breaks · Content-Type Header – The exact Content-Type header to send, including when to add charset=utf-8.
On the related tools: Markdown to HTML – Convert Markdown into clean, safe HTML you can paste anywhere.