On-device AI — professional-grade, fully private.

JSON to CSV Converter

JSON to CSV with your column order and exact numbers kept.

  • Runs on your device
  • Column order and big numbers kept exact
  • One file free · Batch ZIP with Pro
  • Free to 10 MB · Pro to your device's limit

Drop a JSON file here

Up to 10.0 MB · .json, .jsonl, .ndjson, .geojson, .txt

Preview

Paste JSON or drop a file to see the table.

Settings
Rows

Used when the JSON holds more than one list of records.

Columns
Arrays

["a","b"] as one cell, as tags.0 / tags.1, or as one row per item.

Joins nested keys in a header: user.name

Output
Delimiter

Keeps accented and non-Latin text readable when the file is opened in Excel.

Line endings

How to convert JSON to CSV

  1. 1

    Paste or upload JSON

    Paste JSON into the box and press Use this JSON, or drop a .json, .jsonl, .ndjson or .geojson file. UTF-8 and UTF-16 files are read automatically.

  2. 2

    Check the rows and columns

    The preview shows the first 100 rows as a table. If your file is an API response, the record list inside it is found for you; pick another list from Rows come from if there are several.

  3. 3

    Choose the format

    Pick comma, semicolon, tab or pipe, how arrays are written, the key separator, the header row and line endings. Excel-ready adds the byte-order mark Excel needs for accented text.

  4. 4

    Convert and download

    Press Convert to CSV, then download the file or copy it. If you change a setting afterwards, the result is marked out of date and Re-apply brings it up to date.

What this JSON to CSV tool offers

This JSON to CSV converter turns the records in a JSON file into rows and their fields into columns, without reordering, rounding or dropping anything on the way. Paste JSON or upload a file, check the table preview, and download a CSV that opens correctly in Excel, Google Sheets, a database import or a data pipeline.

  • Finds the records for you — A top-level array, an API response that wraps the list in data, results or items, a single object, or JSON Lines with one record per line.
  • Flattens nested objects — Nested keys become columns like address.city or user.profile.name, joined by a separator you choose.
  • Three ways to write arrays — Keep an array as JSON text in one cell, spread it over one column per item, or repeat the record once per item.
  • Keeps your column order — Columns appear in the order their keys first appear in the file, not alphabetically.
  • Keeps numbers exact — Long IDs, trailing zeros and scientific notation are copied character for character.
  • Excel-ready output — A UTF-8 byte-order mark so accents survive, your choice of delimiter and line endings, and quoting wherever a value needs it.
  • Says what it found — Duplicate keys, renamed columns, formula-like cells and very deep nesting are listed above the preview instead of being handled silently.

Which part of the JSON becomes the rows

A CSV is one table, so the first question is which list in the JSON holds the records. The tool answers it from the structure of your file and shows the answer above the preview.

  • An array at the top — Each element is one row. Plain values such as numbers or strings go into a column called value.
  • An API response — In {"status":"ok","data":[…]} the list under data becomes the rows and status is left out. Lists up to three keys deep are found, such as response.data.items.
  • Several lists — The longest list is chosen first. Rows come from lets you pick another list, or the whole document as a single row.
  • A single object — An object with no list of records inside it becomes one row, with one column per field.
  • JSON Lines — A .jsonl or .ndjson file, or any text with one JSON value per line, gives one row per line. A broken line is reported by its line number.

Nested objects and arrays

Nested data has to be laid out flat to fit in a table. Objects always become extra columns; arrays can be written in three ways, and the right one depends on where the CSV is going.

  • Nested objects — Keys are joined with the key separator, so {"user":{"name":"A"}} becomes a user.name column. Change the separator to _ or / if your system expects it.
  • Arrays as JSON text — The default. ["red","blue"] is written into one cell exactly as JSON, so nothing is lost and the row count matches the record count.
  • One column per item — tags.0, tags.1 and so on. Useful for short arrays of a fixed length, such as coordinates.
  • One row per item — The record is repeated once for each item, so an order with three line items becomes three rows. Two arrays in one record are combined, every item with every item.
  • Empty objects and arrays — {} and [] keep their column with an empty cell, so the column exists in every row even when the value is empty.
  • Very deep nesting — Below 64 levels the rest of the value is kept as JSON text in one cell, and the notice above the preview says so.

Column order and numbers, kept exactly

Two things go wrong in most converters without any warning, and both are visible only after the CSV is already in use.

  • Column order — Columns follow the order of the keys in your file. A key first seen in a later record is added after the existing columns, where it first appeared.
  • Number-like keys — Keys such as 2024 or 10 stay where you wrote them. Many tools move them to the front because of how JavaScript stores object keys.
  • Large integers — 12345678901234567890 and 9007199254740993 are copied exactly. A converter that reads numbers as floating point changes them to …567000 and …992.
  • Number spelling — 1.50, -0 and 1e-7 stay exactly as written, so prices keep their trailing zero.
  • Duplicate keys — If an object has the same key twice, the last value is used (as every JSON reader does) and the notice lists where it happened.
  • Two paths, one name — A key called "a.b" and a nested a → b would both be headed a.b. Both columns are kept; the second is named a.b (2) and the notice says so.

Options explained

Every option is free and applies to the whole file. Changing one after converting marks the result out of date until you re-apply.

  • Rows come from — Which list in the JSON becomes the rows. Detected automatically; shown when the file has more than one candidate.
  • Arrays — JSON text in one cell, one column per item, or one row per item.
  • Key separator — The text between nested key names in a column header. A dot by default.
  • Delimiter — Comma, semicolon, tab or pipe between values.
  • Header row — Whether the first line lists the column names. Turn it off when appending to an existing file.
  • Excel-ready — Adds the UTF-8 byte-order mark so Excel reads accented and non-Latin text correctly. Turn it off for tools that do not expect one.
  • Line endings — LF for macOS, Linux and most tools; CRLF for Windows software that requires it.

When to use each delimiter

Match the delimiter to where the CSV will be used.

  • Comma (,) — Default and most common. Use for generic CSV, imports, and most spreadsheets in English locales.
  • Semicolon (;) — Common in European locales where comma is the decimal separator. Use when opening in Excel in such regions.
  • Tab — Use for tab-separated values (TSV) or when pasting into tools that expect tabs.
  • Pipe (|) — Often used in data pipelines and ETL. Use when your downstream system expects pipe-delimited files.

Opening the CSV in Excel

Excel decides how to read a CSV from the first bytes of the file and from your regional settings, which is why the same file can look perfect on one computer and broken on another.

  • Accented characters — Keep Excel-ready on. Without the byte-order mark, Excel on Windows reads UTF-8 as the legacy code page and ü becomes ü.
  • Everything in one column — Your Excel expects a semicolon. Choose Semicolon, or use Data → From Text/CSV and pick the delimiter.
  • Formula-like cells — A value starting with =, +, - or @ may run as a formula when opened. The tool never changes your data, but it lists those cells so you can check them before sharing the file.
  • Long numbers shown as 1.23E+19 — That is Excel's display, not the file. Import the column as Text to see every digit; the CSV holds the exact value.

Why convert JSON to CSV?

CSV is widely supported by spreadsheets, databases, and ETL tools. Converting JSON to CSV lets you open API or app data in Excel, Google Sheets, or import it into other systems.

  • Spreadsheets — Open the CSV in Excel or Google Sheets for viewing, filtering, or sharing with others who don’t work with JSON.
  • Import and ETL — Many databases and data pipelines accept CSV. Flattened columns (e.g. user.name) keep structure in a single table.
  • Reports from API data — Turn an exported API response into a table you can sort, filter and chart without writing code.

How JSON to CSV conversion works

Your file is read in the background, without converting numbers or reordering keys, and kept in memory so changing a setting updates the preview straight away.

  • Read — The encoding is detected, then the JSON is read exactly as written: key order, number text and duplicate keys are all kept. Invalid JSON is reported with its line and column.
  • Find the rows — The record list is located — the top-level array, a list inside an API response, the lines of a JSON Lines file, or the single object.
  • Flatten — Each record's nested keys become columns in first-seen order; arrays are written the way you chose; empty values keep their column.
  • Write — The table is written with your delimiter and line endings, quoting any value that contains the delimiter, a quote or a line break, and the byte-order mark if Excel-ready is on.

Large files are read and converted in the background, so the page stays responsive: about 50,000 records become a CSV in well under a second on a typical laptop.

Frequently asked questions

How does the JSON to CSV converter work?

Paste or upload JSON and the tool finds the records: the elements of a top-level array, the record list inside an API response such as data or results, or a single object as one row. Each record is flattened so nested keys become columns like user.name, the first row lists the columns, and you download the CSV or copy it.

Is my data private and secure?

Yes. The file is read and converted on your device and is never uploaded. Nothing is sent to our servers, so the tool keeps working once the page has loaded, even offline.

What JSON structure is supported and how is nesting handled?

An array of objects such as [{"id":1},{"id":2}], a single object, an API response that wraps the records in a key, JSON Lines with one object per line, and arrays of plain values. Nested objects are flattened with the key separator, so {"user":{"name":"A"}} becomes a column user.name. Arrays are written as JSON text in one cell by default, or as one column per item, or as one row per item.

Which delimiters and encodings can I choose?

The output delimiter can be a comma, semicolon, tab or pipe. The input encoding is detected for you: UTF-8 with or without a byte-order mark, and UTF-16 in either byte order. The CSV is written as UTF-8, with the byte-order mark Excel needs switched on by default.

Is there a file size limit and is it free?

Converting is free for one file at a time up to 10 MB, which is about 50,000 typical records. Pro removes the fixed cap, so the limit is what your device can handle, and converts up to 50 files in one run with a single ZIP download.

Can I use it offline and on mobile?

Yes. The converter works in any modern browser on desktop, tablet or phone, and once the page has loaded it keeps working without an internet connection. Paste JSON or pick a file, convert, then download the CSV or copy it into Excel or Google Sheets.

Why are the columns in the same order as my JSON?

Because that is the order you wrote them in. Each column appears where its key is first seen, record by record, so id, name and email stay in that order. Converters that sort the header alphabetically, or that read the file through a parser that moves number-like keys such as 2024 to the front, reorder your columns without saying so.

Are large numbers like IDs kept exactly?

Yes. Each number is copied exactly as it is written in the file, so 12345678901234567890 stays 12345678901234567890, 1.50 stays 1.50 and 1e-7 stays 1e-7. Converters that turn numbers into floating-point values silently change any integer longer than 15 or 16 digits.

Why does Excel show strange characters, and how do I fix it?

Excel opens a CSV without a byte-order mark using your system's legacy code page, so ü, é or ç turn into two strange characters. Keep Excel-ready switched on and the CSV starts with the UTF-8 byte-order mark, which tells Excel to read it correctly. Use semicolon if your Excel uses a comma as the decimal separator.

What do the warnings above the preview mean?

They name anything about your data that a CSV cannot show directly: keys that appear twice in one object (the last value is used, as every JSON reader does), two different key paths that produce the same column name (the second is renamed with (2)), and cells that start with =, +, - or @, which a spreadsheet may run as a formula. Nothing in your data is changed; the warning tells you so you can decide.

4.8out of 5

from 172 ratings

Rate this tool

Tap a star — it takes a second