JSON to CSV Converter

Why non-English characters look broken in Excel, and the BOM

This is the most common problem users of this tool face. The conversion works fine, but when you open the CSV file in Excel, you see question marks or garbled text like "占쏙옙".

The cause is Excel. On many systems, Excel doesn't assume a file is UTF-8. Instead, it tries to read it using a local legacy encoding (like CP949 in Korea or Windows-1252 in Western Europe). When it tries to interpret a 3-byte UTF-8 character using the wrong character map, the text breaks.

The solution is a BOM (Byte Order Mark). If you prepend three special bytes (EF BB BF, or U+FEFF) to the start of the file, Excel recognizes it as UTF-8. The "Excel-compatible UTF-8 BOM" option does exactly this.

The problem of disappearing leading zeros in phone numbers

It's that classic problem where a phone number like `01012345678` becomes `1012345678` in Excel, or a 16-digit credit card number turns into `1.2346E+15`. This happens because CSV is plain text with no type information, and Excel makes its own guess: "if it looks like a number, I'll treat it as a number."

This tool's CSV → JSON conversion is designed to avoid this pitfall.

• Values starting with a `0` (like phone numbers or postal codes) are not converted to numbers but are kept as strings.
• Numbers with 16 or more digits (like credit card or some account numbers) are also kept as strings. This is because storing them as standard JavaScript numbers would lose precision.
• The text `true`, `false`, and `null` are converted to their respective boolean and null types. If you uncheck "Auto-detect numbers & booleans", all values will remain strings.

Flattening Nested JSON — The Rules of Dot Notation

API responses are often nested, but CSV is a flat table, so the nested structure must be unfolded. When "Flatten nested objects (a.b.c)" is enabled, the tool creates column names by joining the object path with dots.

• For arrays, the index becomes part of the key, like `tags.0` and `tags.1`. If the number of items varies between rows, the number of columns will expand to accommodate the longest array.
• If flattening is disabled, nested objects and arrays are embedded as a JSON string within a single cell. Use this when you need to preserve the original structure.
• In the CSV → JSON direction, enabling flattening does the reverse: it restores column names with dots back into nested objects, allowing for a perfect round trip.
• If keys differ from row to row, the tool collects all unique keys from all rows to form the columns, leaving cells blank where a key doesn't exist for a particular row.

{ "orderId": "A-0042",
  "customer": { "name": "John Doe", "contact": "555-1234-5678" },
  "tags": ["new", "VIP"] }

→ columns: orderId | customer.name | customer.contact | tags.0 | tags.1

Values Containing Quotes, Commas, or Newlines

In CSV, if the delimiter (e.g., a comma) appears inside a value, it will break the table structure. According to the RFC 4180 standard, such values must be enclosed in double-quotes. Any double-quotes within the value itself must be escaped by doubling them. This tool handles all of this automatically.

• Changing the delimiter to a semicolon (;) helps with opening files in versions of Excel that use it as a list separator (common in many European regions).
• If you select Tab (TSV), the saved file extension will be `.tsv`. For data with many commas in the content (like addresses or product descriptions), TSV can be cleaner and easier to read.
• Line endings are written as CRLF (Carriage Return, Line Feed) for maximum compatibility with Excel and other Windows applications, as this is the de facto standard.

Input Formats and Limitations

To create a proper table, the input should be an "array of objects," which provides a clear source for column names. However, since real-world API responses are rarely so clean, the tool intelligently handles a few common variations.

• Wrapped arrays like `{ "data": [ … ] }`: If the root object contains exactly one key whose value is an array, the tool will automatically use that array. If there are multiple keys with array values, it can't know which to use, and the output will be a single-row table of the parent object.
• A single JSON object: This will be converted into a table with one data row.
• Drag-and-drop a `.json`, `.csv`, or `.tsv` file. The tool will automatically detect the file type and set the conversion direction and delimiter. The file is processed entirely in your browser and never uploaded; max file size is 8MB.