Smart Hairstyle Design

Problem Addressed

When guiding a client through a haircut or image refresh, a single portrait photo is often enough to start, but designers still need several visible hairstyle options on the same face. Manual iteration is slow and hard to compare. This skill solves that by taking a clear face image and producing a 3x3 preview grid of nine hairstyle variants, while also keeping the original image and nine split sub-images for review.

How It Works

• Input handling: accepts portrait_image as Base64 or an accessible URL, then runs safety and compliance review before generation.
• Prompt generation: uses the person's appearance and settings such as aspect_ratio and need_text to build hairstyle prompts, constrained to a 3x3 grid layout.
• Async generation: submits the image generation job without blocking, returns a task_id, and polls until the result is ready, which fits longer generation jobs.
• Output split: after the combined image is available, it splits the 3x3 grid into nine independent images and returns the source image, combined image, and sub-image URLs.

Boundaries

Best for clear front-facing portraits that pass review and can be supplied as Base64 or a URL. Real generation depends on ShowAPI credentials; without them, the system may return a simulated preview only. The workflow is asynchronous, so the caller should keep polling and surface each poll response to the user.