Ideogram Sketch to Render API vs a Sume image reference edit

Ideogram's Sketch to Render "turns a fashion sketch into photorealistic garment or product imagery" from an uploaded sketch_image. Sume has no fashion-specific tool in the docs read; the equivalent is an image edit with the sketch URL in image_urls and a prompt that states materials and rendering.

Ideogram facts are from its Sketch to Render page, read 2026-10-01.

What does Sketch to Render take?

POST /v2/tool/sketch-to-render is multipart. sketch_image (up to 50 MB; JPEG, PNG, WEBP, HEIF, AVIF, GIF, BMP, TIFF, MPO) and instruction are required; instruction carries material and rendering direction. Optional fields include aspect_ratio (default 1:1), quality, seed, num_images, and webhook_url. It returns a generation_id to poll.

How do I do it on Sume?

The Image 1.0 page lists "Edit / reference" as prompt plus image_urls (1 to 10 public HTTPS URLs). Put the sketch first, write the fabric, colorway, and lighting in prompt, and set aspect_ratio and num_images (1 to 4) as needed (see Image 1.0). The model decides how closely the sketch's silhouette is kept; nothing in the docs promises it. A close cousin is GPT Image 2.5 sketch to image.

What gets rejected?

Reference URLs must be public HTTPS; localhost, private-network, and non-HTTPS URLs are rejected before submission. A parameter the selected model does not list returns 400 unsupported_parameter instead of being dropped, so do not send fields such as seed unless the model lists them.