Inherits: RefCounted < Object

Helper class for creating and parsing JSON data.


The JSON enables all data types to be converted to and from a JSON string. This useful for serializing data to save to a file or send over the network.

stringify is used to convert any data type into a JSON string.

parse is used to convert any existing JSON data into a Variant that can be used within Godot. If successfully parsed, use data to retrieve the Variant, and use typeof to check if the Variant's type is what you expect. JSON Objects are converted into a Dictionary, but JSON data can be used to store Arrays, numbers, Strings and even just a boolean.


var data_to_send = ["a", "b", "c"]
var json_string = JSON.stringify(data_to_send)
# Save data
# ...
# Retrieve data
var error = json.parse(json_string)
if error == OK:
    var data_received = json.data
    if typeof(data_received) == TYPE_ARRAY:
        print(data_received) # Prints array
        print("Unexpected data")
    print("JSON Parse Error: ", json.get_error_message(), " in ", json_string, " at line ", json.get_error_line())

Alternatively, you can parse string using the static parse_string method, but it doesn't allow to handle errors.

var data = JSON.parse_string(json_string) # Returns null if parsing failed.

Note: Both parse methods do not fully comply with the JSON specification:

  • Trailing commas in arrays or objects are ignored, instead of causing a parser error.

  • New line and tab characters are accepted in string literals, and are treated like their corresponding escape sequences \n and \t.

  • Numbers are parsed using String.to_float which is generally more lax than the JSON specification.

  • Certain errors, such as invalid Unicode sequences, do not cause a parser error. Instead, the string is cleansed and an error is logged to the console.







get_error_line ( ) const


get_error_message ( ) const


parse ( String json_string )


parse_string ( String json_string ) static


stringify ( Variant data, String indent="", bool sort_keys=true, bool full_precision=false ) static

Property Descriptions







Contains the parsed JSON data in Variant form.

Method Descriptions

  • int get_error_line ( ) const

Returns 0 if the last call to parse was successful, or the line number where the parse failed.

  • String get_error_message ( ) const

Returns an empty string if the last call to parse was successful, or the error message if it failed.

Attempts to parse the json_string provided.

Returns an Error. If the parse was successful, it returns @GlobalScope.OK and the result can be retrieved using data. If unsuccessful, use get_error_line and get_error_message for identifying the source of the failure.

Non-static variant of parse_string, if you want custom error handling.

Attempts to parse the json_string provided and returns the parsed data. Returns null if parse failed.

Converts a Variant var to JSON text and returns the result. Useful for serializing data to store or send over the network.

Note: The JSON specification does not define integer or float types, but only a number type. Therefore, converting a Variant to JSON text will convert all numerical values to float types.

Note: If full_precision is true, when stringifying floats, the unreliable digits are stringified in addition to the reliable digits to guarantee exact decoding.

The indent parameter controls if and how something is indented, the string used for this parameter will be used where there should be an indent in the output, even spaces like "   " will work. \t and \n can also be used for a tab indent, or to make a newline for each indent respectively.

Example output:

## JSON.stringify(my_dictionary)

## JSON.stringify(my_dictionary, "\t")
    "name": "my_dictionary",
    "version": "1.0.0",
    "entities": [
            "name": "entity_0",
            "value": "value_0"
            "name": "entity_1",
            "value": "value_1"

## JSON.stringify(my_dictionary, "...")
..."name": "my_dictionary",
..."version": "1.0.0",
..."entities": [
........."name": "entity_0",
........."value": "value_0"
........."name": "entity_1",
........."value": "value_1"