Settings¶
The gear in the studio home's header opens the Weaver settings. Every field here overrides an engine default — a blank field means "use the default," and the default is always visible as the field's placeholder, so you can see exactly what you're changing. Clear a field and save to go back to stock.
These are studio-wide settings: they apply to every build, not per session.
How Much the Studio Does in One Go¶
| Setting | What it controls |
|---|---|
| People per proposal | How many people one Propose people pass suggests on a world's People pane. |
| Questions before a profile is suggested | How many weave answers the pane considers "enough for a solid profile." You can always weave earlier or keep answering past it. |
| Extra questions an interview may ask | The cap on deepening questions after a build's essentials are covered. Raise it for builds where you want a thick depth or lore book out of one session; post-finalize Add lore on a world is uncapped regardless. |
| Most people picked up from one build | The ceiling on how many people the world harvests into the roster from your own interview material at finalize. |
How the Model Runs¶
Two temperatures, split by the kind of work. Higher is looser; lower is more deterministic.
| Setting | What it covers |
|---|---|
| Writing temperature | Everything the model composes for you — questions, directions, the Bible, rendered fields, lore entries, profiles. |
| Judging temperature | Everything the model checks — the quality gates, the read-back, the import reading. |
The split exists because the two jobs want different things: writing benefits from room to move, judging benefits from consistency. If rendered fields feel samey, nudge the writing temperature up; if gate verdicts feel erratic between re-checks, bring the judging temperature down. Blank for both is a sensible place to live.