JSON to CSV

updated 12 March 2026

Before you start

The tool expects one of these three shapes:

• JSON array of objects — the common REST response shape: [{...}, {...}].
• JSON Lines (JSONL / NDJSON) — one JSON object per line, no outer brackets. This is what you get from jq without the -s flag, and from most streaming log pipelines.
• Single object — {...}. It gets wrapped in a one-element array automatically.

The objects don't need to share the same keys. The header row is the union of every key that appears anywhere in the input, in first-seen order. Missing fields become empty cells.

How to use it

1. Paste JSON or drop a .json / .jsonl file into the left pane.
2. Click Convert. The CSV appears on the right and the status bar reports the row count.
3. Use Download .csv to save the result, or copy it from the right pane.

If the input parses as JSONL it's auto-detected — you don't need a toggle. The output is always UTF-8 CSV with CRLF line endings (what Excel expects) and standard RFC 4180 quoting (cells with commas, quotes, or newlines get double-quoted).

Example

Input:

[
  { "id": 1, "name": "Alice", "tags": ["ops", "lead"] },
  { "id": 2, "name": "Bob, Jr.", "address": { "city": "Paris" } }
]

Output:

id,name,tags,address
1,Alice,"[""ops"",""lead""]",
2,"Bob, Jr.",,"{""city"":""Paris""}"

Notice three things: "Bob, Jr." is quoted because it contains a comma; nested objects and arrays are preserved as stringified JSON inside one cell; and the tags / address columns only appear because one record had them — the other gets an empty cell.

Tips & common pitfalls

• Flatten before you convert. If you want address.city as its own column, pre-process the JSON (a one-line jq or map usually does it). This tool intentionally doesn't flatten — doing it wrong silently corrupts data.
• Arrays in cells = JSON strings. An object's tags: ["a","b"] becomes the literal string ["a","b"] inside one CSV cell. If you want one-tag-per-row, duplicate the records upstream.
• Non-object inputs fail. A top-level JSON array of strings (["a","b"]) has nothing that maps to columns. Wrap each value in an object first.
• Keys are emitted in first-seen order. If your first record is missing a field, that field goes toward the right of the header.
• Excel + UTF-8 headaches. The output is UTF-8 without a BOM. If non-ASCII characters look broken when you open the CSV in Excel, open it via Data → From Text/CSV and pick UTF-8 there instead of double-clicking.

Troubleshooting

Frequently asked questions