Series preset
A short reference for the Series Lab preset file: which fields come from which tab. Not the full specification.
One file per series
The preset is presets/<id>.json. The studio fills it in from the tabs. You do not have to open it by hand. The same seed and the same file build the same characters again. Mint and reveal are not in this file: they come from release.json after the deploy kit, through series mint and series reveal.
The ID is latin letters, digits and hyphens. It becomes the file name.
What is stored
| What | Where you set it | Why |
|---|---|---|
| Name, ID, supply, seed | Overview | The series passport |
| Path to the base image | Base | The primary base image |
| Traits and options | Constructor | A snapshot of the pack: name, weight, prompt |
| Prompt template | Constructor | What stays in every description, which traits change, the ending |
| Style lock | Constructor | Angle, light, mood, framing |
| Engine and reference strength | Overview or Generate | Who draws, and how closely to follow the base |
Traits in the file are a copy from the moment you saved. Updating a pack in the library does not change this series.
What the start of the file looks like
{
"schema": "series-preset/v0",
"id": "my-heroes",
"name": "My Heroes",
"seed": "heroes-v1",
"size": 24
}The rest — traits, the template and the style lock — is written when you save the Constructor. The full schema, for people editing JSON next to the code, stays in the repository.