The Design Tokens Format Module is a specification from the Design Tokens Community Group (DTCG) for exchanging design tokens in JSON. It defines how a file represents named tokens, groups, types, values, and references. The group lists version 2025.10 as stable. It is a Community Group specification published outside the W3C Standards Track.
What does a DTCG token look like?
The following 2025.10 example groups two duration tokens under motion. The base duration is 150 milliseconds. The feedback token reuses it by reference.
{
"motion": {
"$type": "duration",
"fast": {
"$value": { "value": 150, "unit": "ms" }
},
"feedback": {
"$value": "{motion.fast}"
}
}
}
The base token motion.fast inherits its duration type from the group. motion.feedback gets its type from the token it references. The curly-brace reference points to the token at motion.fast, so motion.feedback resolves to the same duration value of 150 milliseconds. The motion group organizes the tokens. Each child token has its own $value.
What does the Format Module cover?
The Format Module defines the exchange structure and the rules for interpreting it. It covers supported types, including composite types that combine related properties into one value. The separate Resolver Module describes how to resolve token values for different contexts, such as a light or dark theme.
Use the format to share the data between tools. Your team still decides how to organize tokens, which names communicate a semantic role, and when to change a token. A platform build can turn the resolved values into the files an application uses.
How do you check interoperability?
Pick a small set of real tokens and verify what each tool can import, resolve, and export. Include the types you use and a reference across groups. If the workflow supports themes, check the same role in two contexts. Compare the resolved output with the values expected by the application. Test for broken references and circular references.
Record the supported specification version and any extensions your workflow relies on. A valid JSON file still needs to follow the format’s token rules. If a tool drops a reference or exports a literal instead, check whether the next tool still has the information your workflow needs.