What output resolutions and formats does Ray 3.2 support?

Ray 3.2 outputs native 1080p resolution (540p, 720p, and 1080p options available) with real 16-bit HDR color. It supports EXR export in ACES2065-1 color space, plus standard video formats for input/output.

Can I edit my own video footage with Ray 3.2?

Yes, the Video Edit workflow allows you to upload a clip (MP4, MOV, or WMV), describe the visual change, and Ray 3.2 transforms it while preserving the original motion, timing, and performance. It supports up to 20 seconds at 1080p.

How many keyframes can I use in a single clip?

You can set up to 16 keyframes across a single clip to control exact moments where motion, lighting, or story beats change.

Does Ray 3.2 support HDR workflow and professional color grading?

Yes, it generates native 16-bit HDR color (not post-processed) and exports as EXR in the ACES2065-1 color space, ready for grading in Nuke, DaVinci Resolve, or other VFX pipelines.

What are the input options for images and videos?

For images: JPG, PNG, WEBP. For videos: MP4, MOV, WMV. You can also use text prompts alone or with reference images.

Ray 3.2 AI Video Generator — Generate cinematic AI video with frame-level control

Ray 3.2 is Luma AI's latest video generation model that turns text descriptions and images into fluid, cinematic clips. It handles complex motion, lighting changes, and multi-subject scenes with notable consistency, offering frame-level control and native HDR output. The tool is built for creators, marketers, and storytellers who want professional video without expensive production, providing a fast, reliable starting point for demos, concept visualization, or content at scale.

Ray 3.2 operates directly in the browser with four core workflows: Text to Video, Image to Video, Video Edit, and Video Reframing. It supports up to 16 keyframes per clip, real 16-bit HDR generation, and EXR export in ACES2065-1 color space, making it suitable for VFX pipelines. The model preserves motion, structure, and timing when editing existing footage, and it is available through platforms like Adobe, Freepik, Fal.ai, and others.