Online CSV/JSON Converter: Convert Between Table Data Formats

What Are CSV and JSON

CSV (Comma-Separated Values) is tabular text: a header row, then one record per row, fields separated by commas by default.

JSON expresses objects and arrays for APIs and programs. ToolVault's CSV/JSON Converter converts both ways in the browser — nothing is uploaded.

Comparison example (matches this site's output)

CSV:

name,email,age
Alice,alice@example.com,28
Bob,bob@example.com,32

JSON (all strings — this tool does not infer types):

[
  { "name": "Alice", "email": "alice@example.com", "age": "28" },
  { "name": "Bob", "email": "bob@example.com", "age": "32" }
]

Note age stays "28", not the number 28. Coerce per column downstream when you need numeric types.

CSV vs JSON

How to use this converter (real UI)

Open the CSV/JSON Converter:

1. Paste CSV (with a header) or a JSON array of objects on the left.
2. Click CSV → JSON or JSON → CSV — the tool does not auto-detect direction and does not sniff delimiters.
3. Choose a delimiter: comma ,, semicolon ;, or Tab (shared for parse and emit).
4. Copy or download the result.

Implementer's notes (csvToJson / jsonToCsv)

Delimiter

• UI offers , / ; / Tab; the second argument to csvToJson / jsonToCsv matches the control.
• Locale Excel exports that use ; can be pasted directly with the semicolon option—no need to pre-normalize via CSV/TSV.

Quoting (RFC 4180 subset)

• Fields with commas, quotes, or newlines must be wrapped in "...".
• Escape quotes by doubling: "said ""hi""" → said "hi".
• A " mid-field (not starting the field) → CSV_QUOTE_POSITION_ERROR.
• Unclosed quotes → CSV_QUOTE_NOT_CLOSED.
• Newlines inside quotes are legal (one logical record can span physical lines).

Headers

• Row 0 is the header; keys are trimmed.
• Header-only input → [].
• Blank input → [].

Empty cells and column counts

• Empty cells become "", never JSON null.
• Every data row must match the header width or you get CSV_FIELD_COUNT_MISMATCH.
• Fully blank rows are skipped.

Newlines and BOM

• \n and \r\n are accepted (\r alone is skipped).
• The whole input is trim()'d; JS trim strips a leading UTF-8 BOM (U+FEFF), so Excel "CSV UTF-8" headers usually do not become BOM+name.
• If another toolchain keeps the BOM on the first header, that path did not strip it — this one does.

JSON → CSV extras

• Input must be an array of objects.
• Headers are the union of keys (first-seen order); missing keys become empty cells.
• null / undefined → empty cell.
• Values starting with =, +, -, @, etc. get a leading tab to reduce Excel/WPS formula injection risk.
• Fields with ,, ", or newlines are RFC-quoted.

Large-file boundary

• The full string is held in memory; fine for interactive and mid-size tables.
• Hundreds of thousands of rows or ~100MB+ blobs belong in streaming CLIs or database loaders — not a tab.

RFC 4180, Excel dialects, and the real cracks

RFC 4180 describes a common shape (comma + quotes + CRLF) but is not globally enforced:

• Delimiter follows locale — German Excel often emits ;; this tool stays on ,.
• Quoting is "when needed" — exporters that skip quotes corrupt the first important address field.
• Record ≠ physical line — quoted newlines break naive split('\n') scripts.
• BOM — decides whether the first header join fails silently.

Treat CSV as a dialected text table, not "JSON with commas".

Round-trips are not identity

Even staying on this site:

• Types stay strings end-to-end (predictable — no silent 00123 → 123).
• Column order after JSON→CSV follows key-union scan order, which may differ from the original CSV.
• CSV has no null; the return trip will not invent one.
• Nested objects are not flattened here. For dotted keys and custom delimiters, use JSON → CSV.

Cluster links: JSON, SQL, YAML

Related posts: JSON/CSV guide, CSV to SQL, YAML/JSON.

Failure cheat sheet

FAQ

Commas inside fields?

Wrap in quotes: "Shanghai, Pudong". Or switch to TSV via CSV/TSV.

Why are numbers still strings?

CSV→JSON here deliberately skips type inference so order IDs and leading zeros survive. Coerce in the app or with SQL CAST.

Difference from the JSON→CSV tool?

• CSV/JSON: both directions, fixed comma, string values, strict column counts.
• JSON→CSV: JSON-first export, optional flatten, custom delimiter, header toggle.

Large files?

Interactive sizes yes; 100MB+ use a local pipeline. Browser memory roughly doubles (source string + object graph).

From ToolVault. Related: CSV/JSON, JSON→CSV, CSV→SQL, YAML↔JSON. More on the homepage.