Input CSV
Showing
Node Details
Select a node to view details
Turn CSV rows into JSON arrays of objects — quoted fields, custom delimiters and types handled for you
Select a node to view details
Format, validate, and explore JSON — 100% client-side, private by design
Paste XML, get clean JSON — attributes, repeated elements and text nodes handled correctly
Turn CSV rows into JSON arrays of objects — quoted fields, custom delimiters and types handled for you
Flatten nested JSON into clean CSV columns — download Excel-ready .csv files in one click
Compress JSON to a single line and watch exactly how many bytes you save
Compare two JSON documents semantically — key order and whitespace never count as differences
Escape JSON strings for embedding — or unescape stringified logs back into readable JSON
One-click repair for broken JSON, JSON5, JS objects, Python dicts and truncated LLM output.
Turn any JSON payload into clean TypeScript interfaces or type aliases — merge arrays, detect optional fields and unions.
Parse YAML configs — K8s, Docker Compose, GitHub Actions — into clean JSON, then explore and edit visually.
Generate clean, idiomatic YAML from any JSON — smart string quoting, compact list style, 2 or 4 space indent.
Serialize JSON to well-formed XML — @_ attributes, repeated elements, text nodes and safe tag names.
Infer a standards-compliant JSON Schema from example data — formats, required keys, unions and additionalProperties control.
Query any JSON with live JSONPath evaluation — wildcards, recursive descent, slices and filters, all in your browser.
A real character-level state machine — quoted fields containing commas, double-quote escapes (""), and embedded newlines all parse correctly.
Comma, semicolon, tab, and pipe delimiters are detected automatically. European CSVs with semicolons and TSV files just work.
With "First row is header" you get an array of objects keyed by column names. Turn it off for a plain two-dimensional array.
Converted rows land directly in the table view — check the mapping visually and even edit cells inline before copying the JSON.
Drop CSV text into the input panel or upload a .csv/.tsv file. The delimiter (comma, semicolon, tab, pipe) is detected automatically.
Choose whether the first row is a header and whether numbers and booleans should become native JSON types.
Click "Convert to JSON" — results appear in the table view by default. Copy the JSON or download it as a .json file.
How formatjson.net compares to traditional server-based JSON formatters and converters
| Capability / Standard | formatjson.net (Client-Side Workbench) | Traditional Server-Based Tools |
|---|---|---|
| Data Processing & Privacy | 100% local in-browser JavaScript — payloads never leave your device | Often uploads JSON/XML/CSV to remote backend servers |
| Confidential API Keys, JWTs & PII | Safe by design — zero network transmission, works offline via Service Worker | Risk of server access logs or third-party data retention |
| Broken & Truncated LLM JSON Repair | 1-click lexical state machine fixes comments, single quotes, trailing commas, Python literals & unclosed brackets | Fails with syntax error on non-strict JSON |
| 64-Bit Snowflake IDs & Leading Zeros | Preserves unsafe integers (> 2^53 - 1) and leading-zero codes ("00123") without precision loss | Silently truncates 64-bit IDs or strips leading zeros |
| Deep Nested Array Exploration | Interactive spreadsheet drill-down ([n items] →) with breadcrumbs, 3-state sorting & inline cell editing | Flat root-only tables or read-only trees |
| Large File Responsiveness | Chunked lazy tree (200 nodes/batch) + paginated table + 150 KB highlight guard | Main-thread DOM freeze on multi-megabyte files |
Keep "First row is header" checked (it is by default). Each data row becomes a JSON object whose keys are the column names — e.g. id,name\n1,Alice becomes [{"id":1,"name":"Alice"}].
Yes. Auto-detect recognizes commas, semicolons, tabs, and pipes, and you can also pin a specific delimiter manually — handy for semicolon-delimited European CSVs and TSV exports.
The parser follows RFC 4180: a field wrapped in double quotes may contain delimiters, line breaks, and escaped quotes (""). "Smith, John" stays a single cell instead of splitting at the comma.
Yes — "123" becomes 123, "true" becomes true, and "null" becomes null when "Parse numbers & booleans" is on. Uncheck it to keep every value as a string.
Absolutely. The tool opens the result in the interactive table view so you can verify every column. Values can also be edited inline in the details panel before you copy or download the JSON.