Lora Visual

Create raster visuals in one shared cute Japanese-anime editorial language. Choose one initial output mode before generating. Do not run both production paths in parallel. Mode A may switch once to Mode B only after reaching the text-repair stopping condition below.

Use assets/lora-character-reference-v2.png as the canonical character reference whenever the image tool supports references. It is deliberately free of text, logos, labels, and pseudo-text.

Choose the output mode

Mode A — explanatory image

Use when the image must explain a concept, mechanism, workflow, comparison, or tradeoff by itself.

• Deliver a complete PNG or WebP with a finished off-white scene.
• Generate every essential title and label directly inside the bitmap.
• Make the relation visible through objects, paths, states, or repeated materials; labels identify the evidence but do not replace it.
• Do not generate an unlabeled base and add essential words in a separate rendering step.

Mode B — transparent illustration