Image interpretation and weave

Build a base description, independently refine zoom evidence, then generate layer 6 or 7.

Restoring image weave workspace…

Standard images

The first image is primary. Further uploads or single gallery selections are added as standard full-image views.

Choose existing figure description or gallery (layers 1–3)

Existing figure descriptions and galleries

Choose a canonical description, a stored image caption, or an image to interpret again.

No catalog figures found.

Upload a fresh image (layer 5)

Upload source image

(drag and drop)

Choose ImageKit or recent database image (layer 4)

Also include figure-catalog-zoom imagesAlso include reference images

Choose existing ImageKit images

Select in order; the first image is primary.

Up to 1 images per batch.

Save current image set as a new figure

Creates a figure from the primary image, 0 additional standard images, and 0 zoom crops, then embeds its description and interpreted features.

Figure name

Models

Visual interpretation and final weave generation are selected independently.

Visual interpretation model

Weave model

Zoom crop descriptions

Review each crop as a separate editable layer. Chat refinements update only that crop before it is used in a weave.

Assign existing image as zoom crop layer

No zoom crop descriptions yet

Click the source image, zoom in, bound a crop, and submit it for interpretation.

Primary image description

Stored text, direct edits, interpretations, and chat revisions share one non-destructive history (0 revisions).

Lower-body focus anchors

0 detected

Automatically detected anchors are limited to legs, feet, footwear, and related adornments. Only selected anchors guide generation; source subject, pose, scene, and unrelated details are excluded.

No lower-body, footwear, or adornment anchors were detected in the current descriptions.

Weave generation

Layer 6 uses the primary description, selected standard-image descriptions, and selected crops. Layer 7 uses selected crops alone. Active sources rotate and form light composites throughout the output.

CompositionPrompt familyStyle

Layer 6 needs a base description. Layer 7 needs at least one selected crop description.

Output

Generated output will appear here.