Photo to Fun Illustration

Problem

Real photos are often plain to share and lack a stylized, memorable presentation. This skill turns a local photo into an illustration and places it on an Ardot canvas for preview, framing, and sharing.

How It Works

The workflow uses ImageGen image-to-image plus Ardot canvas compositing. It reads a local image path, then generates either a cute chibi style or an ugly-cute doodle style based on user choice. Single-style runs produce one image; dual-style runs use serial generation to reduce timeout risk. input_fidelity controls detail preservation: medium keeps identity for chibi style, while low gives the doodle style more room for loose, exaggerated lines. The skill then creates an Ardot canvas, adds a frame with batch_edit, inserts the image with upload_images, optionally places a message card, and validates layout with capture_screenshot.

Boundaries

Input should be a local absolute path, and large camera files should be resized first to avoid oversized request bodies. Prompts should constrain head-to-body ratio, keep expressions friendly, and avoid mean-spirited caricature. Sports uniforms, team badges, or other sensitive elements may trigger review, so describe them as cartoon characters.