Generating Portraits

Open the portrait modal to generate an AI portrait, upload your own image, pick an aspect ratio, and browse every version. The prompt is built from your design’s style plus the character’s description, which you can edit or auto-write.

Opening the Portrait Modal

Click a character’s portrait thumbnail to open the portrait modal. The modal shows the character’s current portrait (if one exists) on the left and the generation controls on the right.

Prefer to skip the modal? The Generate / Regenerate button on each row kicks off a portrait straight from the row using the design’s current settings — and if the character has no description yet, it auto-writes one first.

Modal Layout

The modal is split into two columns:

Left Column

The portrait preview, which doubles as a drag-and-drop and clipboard-paste target. When more than one version exists, a thumbnail strip sits below it.

Right Column

The segmented Image Prompt, the aspect-ratio picker, and the Generate and Upload Image buttons, with an estimated credit cost shown just beneath.

The Image Prompt

The prompt is assembled from labelled segments so you can see exactly what goes into the image:

Style — The visual style pulled from your active production design. It’s read-only here; change it in the Designs tab and every portrait follows.
Framing — The portrait framing template (e.g. a head-and-shoulders portrait). Also read-only, and managed in Designs.
Description — The editable part. It’s pre-filled with the character’s description; type to refine it, or use the Auto-describe link (labelled Re-describe once there’s text) to regenerate it from script context.

Edits you make to the Description are saved back onto the character when you generate, so the row and the prompt stay in sync.

The model, seed, and any advanced parameters come from your active production design’s defaults rather than from this modal — adjust them in the Designs tab to change them across the whole cast.

Aspect Ratio

When the design’s image model supports multiple aspect ratios, they appear as selectable chips (for example 1:1, 2:3, 3:4). Portraits default to a square 1:1. Pick the ratio you want before generating.

Generate vs Regenerate

If the character has no portrait yet, the button reads Generate; once one exists it reads Regenerate. Each run creates a new version rather than overwriting — earlier portraits stay available in the version strip. The estimated credit cost is shown beneath the button before you commit.

Uploading Your Own Image

Already have artwork or a reference shot? Add it as the portrait in any of three ways:

Upload Image — Click the button to pick a file from your device
Drag-and-drop — Drop an image file onto the preview area; a highlighted “Drop image here” overlay confirms the target
Clipboard paste — Copy an image and press Ctrl+V / Cmd+V with the modal open

An uploaded image becomes a version just like a generated one, so you can switch between uploads and AI results freely.

Browsing Versions

Every generated or uploaded portrait is kept as a version. When there’s more than one:

Chevron arrows overlaid on the image step to older or newer versions
A thumbnail strip below the image (labelled v1, v2, …) lets you jump straight to any version
The model used for the version you’re viewing is shown above the prompt
The save icon in the corner of the image sets the version you’re viewing as the character’s active portrait