Luma Ray 3.2 Image to Video

Luma Ray 3.2 Image to Video brings your still images to life, transforming a single source image into fluid, cinematic motion guided by a text prompt. Rather than generating video from scratch, this model takes an image you already love — a photograph, an illustration, a rendered scene, a character design — and animates it while preserving the look, composition, and feel of that starting frame. The result is motion that feels intentional and grounded, keeping the visual identity of your original artwork intact as it moves.

At its core, the model works by using your image as the first frame of the video, then interpreting your text prompt to decide how the scene should move and evolve. You describe the motion or atmosphere you want — a low-angle shot of a tiger prowling through snow, a slow camera drift, wind moving through a scene — and Ray 3.2 animates from there. This makes it a natural fit for creators who care about consistency: the animation grows out of your image instead of reinventing it.

The model gives you meaningful creative control over how the final video looks. You can choose from a range of aspect ratios to match wherever your work will live, including widescreen 16:9, vertical 9:16 for social and mobile, square 1:1, portrait 3:4, landscape 4:3, and ultra-wide cinematic 21:9. Resolution is adjustable across 540p, 720p, and 1080p, letting you balance detail against your needs — start light for quick tests and step up to 1080p for finished work. Standard clips run five seconds, giving you a focused, shareable motion piece.

One of the most useful features is the ability to guide the animation between two images. By setting both a starting frame and an ending frame, the model interpolates smoothly between them — a powerful way to control exactly where your shot begins and ends. This is ideal for transformations, reveals, and transitions where you know both the first and last beat of the moment you want to capture.

For creators who need more precise choreography, the model supports a multi-keyframe mode. Instead of a single start image, you can pin a series of guide images at specific points along the timeline, giving the model anchors to move through. This unlocks longer ten-second durations and HDR output, and it's the mode to reach for when you want tight control over how a sequence unfolds beat by beat. Note that the multi-keyframe approach is separate from the single-image and start-to-end-frame modes — you choose one anchoring approach per generation.

Ray 3.2 also offers seamless looping for standard clips that don't use an end frame or keyframes. This is perfect for ambient backgrounds, animated wallpapers, social loops, and any piece meant to play continuously without a visible seam. For higher-end delivery, the model can produce HDR video at 720p or 1080p, and can additionally export an EXR sidecar file alongside the MP4 for professional compositing and grading work. HDR and looping can't be combined, so choose the finish that fits your project.

Who benefits? Filmmakers and motion designers can quickly turn concept art, storyboards, or key frames into moving shots for previsualization and mood pieces. Social content creators can animate photos and graphics into eye-catching vertical or square clips. Illustrators and digital artists can see their static pieces move without losing the style they worked hard to establish. Marketers and brand designers can produce short, polished motion assets and loops from existing imagery. Because the animation stays faithful to your source image, it's especially valuable for anyone who needs the motion to respect a specific look rather than drift away from it.

The model handles stylized and transformative motion well, and includes support for lip-sync scenarios, making it useful for character-driven and talking-subject content. Whether your source is photoreal or heavily stylized, Ray 3.2 aims to carry that aesthetic into motion.

A few practical considerations help you get the most from it. The ten-second duration is only available through the multi-keyframe mode — single-image and start-to-end animations are capped at the standard five-second length. Seamless looping only works with standard-range clips that don't include an end frame or keyframes. HDR output requires 720p or 1080p and can't be looped, and EXR export is tied to HDR. Keeping these combinations in mind lets you plan your shot around the finish you need. As with any prompt-guided motion tool, a clear, descriptive prompt that names the camera angle, subject action, and atmosphere will give you the most controllable and cinematic results.

In short, Luma Ray 3.2 Image to Video is a flexible bridge from still to motion — one that respects your original frame, gives you real control over framing, timing, and finish, and scales from quick social loops to HDR deliverables ready for a finishing workflow.