What this converter does
Use JSON → CSV to turn a whole-document JSON array or object into CSV. Use CSV → JSON to convert CSV or TSV files back into JSON. The page parses files, offers controls for nested data and arrays, previews the result, and processes the work locally in your browser.
Convert JSON → CSV
- Drop one or more JSON files into the drop zone or click browse files. The converter detects JSON input automatically.
- If the JSON is invalid, read the reported line-and-column error and fix the source file.
- For an object that contains an array of records, choose the array in Records when the detected choice is not the one you need.
- Set the options below; the preview updates automatically as you change them.
- Download each CSV from its file entry, or click Download all as ZIP.
Choose JSON → CSV options
- Use Delimiter for comma, semicolon, tab, pipe, or a one-character custom value. Include header row and Quote all fields control the CSV output.
- Leave Flatten nested objects enabled to turn nested paths into columns such as
address.city. Key separator supports dot, underscore, and slash:address.city,address_city, oraddress/city. When flattening is disabled, nested objects stay as values until CSV fields are built and then follow JavaScript's standardString(value)conversion; they are not serialized as JSON text. - Under Array handling, choose Join values in one cell, Expand into columns, or Expand into rows. Joined items use semicolons; column expansion creates indexed fields such as
tags.0. - Use the field checklist to include only the columns you need. Select all and Select none change the checklist in one click.
Convert CSV → JSON
- Drop CSV or TSV files into the drop zone or click browse files. The converter detects tabular input automatically.
- Choose the input Delimiter and decide whether First row is a header. Without a header, columns are named
col1,col2, and so on. - Keep Infer numbers, booleans & null enabled to convert recognized values to JSON types. Turn it off to keep every value as a string.
- To rebuild nested objects, enable Un-flatten column names and choose the matching Un-flatten separator: dot, underscore, or slash.
- Choose JSON array of objects or Object keyed by a column. For the keyed shape, select the Key column.
- Review warnings and the automatically updated JSON preview. Download files individually or with Download all as ZIP.
Preview, warnings, and output limits
The CSV preview shows up to the first 200 rows. The JSON preview shows up to the first 50 array records. For JSON → CSV, each file is capped at 20,000 rows and 200 columns; a warning appears when either cap truncates the result. These confirmed caps apply to JSON → CSV, not automatically to CSV → JSON.
CSV warnings identify ragged rows and unterminated quoted fields with line information. Conversion can still produce output, so review warnings before downloading. Remove an unwanted file from the file list with its visible remove control.
Privacy and local processing
Parsing, flattening, un-flattening, previews, and generated output run in your browser. The page describes this as local processing, so the files are not uploaded as part of this workflow.
Frequently asked questions
Does converting between CSV and JSON upload my file anywhere?
No. Parsing and building the output both happen entirely in your browser, in either direction. Your file never leaves your device.
How are nested objects and arrays handled going from JSON to CSV?
Nested objects are flattened into dotted, underscore, or slash-separated column names like address.city, address_city, or address/city. Arrays can be joined into one cell separated by semicolons, expanded into indexed columns like tags.0 and tags.1, or expanded into separate rows so each array item gets its own line.
Can I turn flattened CSV columns back into nested JSON?
Yes. In CSV to JSON mode, turn on Un-flatten column names and pick the same separator that was used to flatten - dot, underscore, or slash column names like address.city are rebuilt into nested objects.
What happens if my JSON or CSV is invalid?
For JSON input, the tool shows a plain-language error naming the problem and its line and column instead of a raw browser stack trace. For CSV input, malformed or ragged rows are reported individually so you can find and fix the issue in your source file.