Sketch to Image

Luma expanded Luma Agents with sketch-to-render and brand-system generation demos, showing rough references turned into finished visuals and branded asset systems. The release matters because style, character and branding control are being packaged into one agent flow instead of separate generation steps.

Threads outlined multi-step pipelines for turning sketches into photoreal scenes, placing furniture with red-box prompts, and swapping in a real face after generation. Prompt-only inpainting still misses precise placement, so creators are using masks, ControlNet, and cleanup passes for tighter control.

Luma is rolling out Uni-1 as a reference-driven image model built around intelligence, directability and cultural taste, with examples spanning sketch conversion and multi-image blends. Use it when references matter more than giant text prompts.

Luma launched Uni-1 and says it can reason through prompts while generating images. Creators report stronger composition on first pass for sketch-to-photo, multiview characters, and reference-led scenes, which should cut correction loops.

Creators showed Kling 3.0 turning sketches into motion, animating ogres and monster fights, and looping branded UI scenes inside node workflows. Try it as a bridge from rough boards to presentable motion tests.