Ray is a video model from Luma AI, the creator of Dream Machine. Version 3.2 puts the focus on visuals: camera movement, lighting, and texture that look like they came straight out of a film or a commercial.
In "Photo to Video" mode, the model takes your image as the first frame and brings it to life while preserving the original composition and aspect ratio. In "Text to Video" mode, it shoots a scene from your description in one of six formats: 16:9, 9:16, 1:1, 4:3, 3:4, or cinematic 21:9. Clips run 5 or 10 seconds at 720p resolution.
Luma Ray 3.2 generates video without sound. If your clip needs footsteps, engine noise, waves, or music, upload it to the "Sound for Video" tool: the AI will match sounds to the footage or compose a track for your edit.