Make-A-Scene

Overview

Make-A-Scene by Meta is a multimodal generative AI system that gives creators precise control over AI-generated images. Instead of relying only on text prompts, you can combine written descriptions with rough sketches or layouts, guiding the model to focus on composition, perspective, and key elements. The system understands your drawing as a scene blueprint and uses it together with your text to generate high-quality, coherent visuals. Designed for artists, designers, storytellers, and anyone exploring visual ideation, Make-A-Scene enables rapid experimentation with different styles and scene variations. You can fix where subjects appear in the frame, define backgrounds, and emphasize important details, while the model fills in textures, lighting, and realistic rendering. This makes it especially useful for storyboards, concept art, character design, and visual prototyping. Built on Meta’s research in multimodal AI, Make-A-Scene aims to make image generation more collaborative, where human intent and machine creativity work together. The tool is free to use and continues to evolve as the underlying models improve. Whether you’re planning a complex scene or simply exploring visual ideas, Make-A-Scene helps turn quick sketches and short text prompts into vivid, production-ready images with far more control than text-only systems.