Free online developer tools
Free online JSON and data tools that run in your browser. Format and validate JSON, and convert between JSON, CSV, XML, YAML, and Markdown. No sign-up, no installs, and your data is processed locally in your browser.
Format and beautify JSON instantly.
Validate JSON structure and fix errors.
Convert JSON data to CSV format.
Convert CSV files to JSON data.
Convert JSON data into XML.
Convert XML into JSON format.
Convert JSON to YAML format.
Convert YAML to JSON instantly.
Flatten nested JSON structures.
Escape JSON strings safely.
Query JSON data using JSONPath.
Clean and format CSV data.
Beautify and format XML.
Convert markdown into HTML.
Converting between JSON, CSV, XML, and YAML is rarely lossless. Knowing where each conversion drops information saves you from discovering it in production.
JSON to CSV loses structure. CSV is a flat grid; JSON is a tree. Any nested
object or array has to be flattened into columns like address.city, or serialised
back into a string inside a cell. CSV also has no type system — every value arrives as text, so
"1", "true", and "null" come back as strings and a
leading-zero postcode like 01234 gets mangled into the number 1234 by whatever
reads it. Use CSV when the destination is a spreadsheet, not when it is another program.
YAML is a superset of JSON, so every valid JSON document is already valid
YAML. Going the other way is where it gets interesting. YAML's implicit typing is famously
surprising: in the YAML 1.1 spec that many parsers still follow, the unquoted value
no parses as the boolean false, which is why the country code for
Norway has broken so many config files that the behaviour has a nickname — the Norway problem.
Unquoted 1.0 becomes a float, and 08 can error as an invalid octal.
If a YAML value must stay a string, quote it.
XML distinguishes attributes from elements; JSON does not. Every
JSON-to-XML converter has to invent a convention for that round trip, commonly prefixing
attribute keys with @ and putting text content under something like
#text. Two converters will not necessarily agree, so an XML document that survives
a round trip through one library may not survive another.
Most invalid JSON fails for one of four reasons, and the error message usually points a few characters past the real culprit rather than at it.
{'a': 1} is a JavaScript object, not JSON.
{a: 1} is likewise JavaScript, not JSON.
\n and \t. This is what a JSON escaper is for.
One thing a validator will not catch: JSON numbers have no precision limit in the spec, but
JSON.parse maps them onto IEEE-754 doubles. Any integer above 253 — a
Twitter/X status ID, a Discord snowflake, a large database key — loses precision silently. If
an API returns large integer IDs, it should send them as strings, and if it doesn't, you cannot
fix it in the browser after the fact.
Usually a trailing comma after the final array element or object property, single
quotes instead of double quotes, or an unquoted key. All three are legal JavaScript and
illegal JSON, so code that was copied out of a .js file often fails.
Parser errors typically report the position just after the problem, not at it.
No. Every JSON, CSV, XML, and YAML tool on this page parses and converts your input in your browser using JavaScript. Nothing is transmitted to ToolzYard, which matters if you are debugging a payload that contains production data or credentials.
JavaScript parses JSON numbers as IEEE-754 doubles, which represent integers exactly only up to 253 − 1 (9,007,199,254,740,991). Anything larger — snowflake IDs, big primary keys — is rounded on parse. The fix belongs on the API side: send such values as JSON strings.
YAML 1.1 implicit typing treats unquoted no, yes,
on, and off as booleans. It is the reason Norway's country
code NO famously breaks config files. Quote any string value whose type
must survive the round trip.
Only by flattening it first. CSV is a flat grid with no concept of nesting, so nested
objects become dotted column names such as user.address.city, and arrays
either explode into multiple columns or get serialised into a single cell. Use the JSON
Flatten tool before converting if you want control over how that happens.