Some newer models reject max_tokens — if so, this switches automatically on retry.
No system prompt here — each wizard step has its own, editable on the step itself.
Settings are stored only in your browser, saved separately per provider. Your API key never leaves this device except to call the provider directly.
{{ wizFootNote }}
Initial InputYour idea or rough prompt — the AI suggests a structure from it
Persona
Structure
Suggested Structure
▤
Describe your idea on the left, then press ✦ Generate Structure. The suggested sections appear here — rename, reorder, or deselect them before writing content.
{{ wt.title }}{{ wt.desc }}
Unchecked sections are skipped. Each checked section becomes its own wizard step.
## Section · ### Subsection — one per line, each optionally followed by a plain-text description line. Edits sync with the tree; matching titles keep their generated content.
Section
{{ wizSecTitle }}
{{ wizSecDesc }}
{{ sb.t }}
Persona · this section
Length / Depth
Extra instructions / context
Shared by all section-content steps.
Section Content{{ wizSecStatus }}
✦
Nothing generated yet. Adjust persona, length and extra instructions, then press ✦ Generate Section.
{{ wb.el }}
Regenerate with feedback
Assembled Prompt{{ wizMissingNote }}
Deselected sections and unchecked blocks are left out. To change a section, click its step above and regenerate or edit it there.
Some newer models reject max_tokens — if so, this switches automatically on retry.
Used only by this tab. The Structure, Persona and Mode instructions are appended after it — including the [option name](ph:choice 1|choice 2) output syntax above.
Settings are stored only in your browser, saved separately per provider. Your API key never leaves this device except to call the provider directly.
System (in AI Settings) + Structure + Persona + Mode · full prompt is logged to the console
Input PromptRough idea in — structured, reusable prompt out