What it does
Luma builds AI video and image models and an agent-based creative workspace around them. The site describes Luma Agents that research, generate and refine work across video, image, audio and text while keeping a brand consistent. Ray 3.2 is the current video model, aimed at giving control over every frame, holding continuity across cuts and finishing in cinematic quality. Uni-1 is a model built around brand awareness, carrying style references, character references and visual grounding into each asset. Ray 3.2 and Uni-1.1 are both announced with APIs. Luma Skills lets you build a creative workflow once and run it again. The company also runs an open physical AI research lab and publishes engineering write-ups on its models. The official site presents the product under the Luma name and offers a 'Try for free' entry point. Plan prices sit on a separate pricing page and are not stated on the home page.
Capabilities
- Ray 3.2 video generation model with frame-level control
- Continuity held across cuts in generated video
- Luma Agents that work across video, image, audio and text
- Style and character references carried through assets
- Luma Skills for reusable creative workflows
- API access to Ray 3.2 and Uni-1.1
When to use it
- You want to generate video from text or from an image
- You need many variants of marketing or social creative quickly
- Your team wants brand-consistent output across media types
- You are a developer who wants video or image generation through an API
