Free YAML to JSON converter

YAML is the format used to write down how software should run: service addresses, server counts, environment variables. The name is an acronym that expands into itself — “YAML Ain't Markup Language”[1]. Indentation carries the structure: whatever sits deeper belongs to the key above it. A single entry is a name, a colon and a value, such as port: 8080.

JSON (JavaScript Object Notation) carries data between programs, and it is the shape APIs answer in. RFC 8259 from the IETF (Internet Engineering Task Force) allows six kinds of value: object, array, string, number, boolean and null[2]. Curly braces open an object, square brackets open an array, and every key name sits in double quotes. XML held that role earlier, and data from an older interface goes through the XML to JSON converter.

Release 1.2 of the YAML specification set itself the goal of making YAML a strict superset of JSON[1]. In practice that means you can paste a JSON file into the left box of this converter and the tool reads it the same way it reads YAML, returning the same data indented with two spaces.

Paste your YAML into the left box and click Convert. JSON indented with two spaces appears on the right. The entry port: 8080 comes back as {"port": 8080}, each further level of indentation becomes a nested object, and a dash list turns into an array.

To take a file from disk, use the Add YAML file link above the input box; it accepts .yaml and .yml. Move the result on with Copy or save it with Download, which names the file converted.json. Clear all empties both boxes. Below the tool sits a source and target format picker leading to neighbouring pairs — the CSV to JSON converter, for instance, when the data arrives from a spreadsheet.

The conversion runs in your browser, so the configuration file stays on your computer. That matters with this format, because YAML files hold access keys and database passwords alongside service names.

The converter reads data according to the core schema from release 1.2 of the YAML specification, which the specification itself calls the recommended default[1]. Only true, True, TRUE, false, False and FALSE are booleans there. The words yes, no, on and off behave like ordinary text and reach the result in quotes. The full set is collected in the table of boolean spellings on this page.

The same rule covers numbers and settles three spellings. The first is a leading zero: 01234 gives the number 1234 in base ten, because base eight has its own 0o prefix in this release. The second is a version number: written without quotes, 1.10 is a number and comes back as 1.1, so write values of that kind in quotes in the source file. The third is a date: the core schema has no date type, so 2024-01-15 reaches the result as the text "2024-01-15", written exactly as in the file.

An anchor and an alias, the & and * spellings, are expanded: every reference gets a full copy of the data in the result. The merge key << works the same way, so a block inheriting from a base configuration holds the complete set of fields in JSON, including the ones it overrode itself. Comments opened with # stay in the source file: RFC 8259 lists six kinds of JSON value and a comment sits outside that list[2].

A file holding several documents is handled separately. The Kubernetes documentation describes grouping multiple resources in one file, separated by a --- line[3], and that is the shape a manifest bundle and the output of helm template take. The converter reads each such document on its own and collects them into a single JSON array, in file order. An empty document after a closing --- line is skipped, so a file with one resource and a trailing separator still gives a single object.

The result is ready for the next tool in line. A list of records goes into a spreadsheet through the JSON to CSV converter, and the way back, from JSON to a readable configuration, is handled by the JSON to YAML converter. When the system on the other side takes XML only, the JSON to XML converter prepares the data.