Luma (Dream Machine)

What Luma (Dream Machine) is

Luma AI is a Palo Alto-based generative AI company founded in 2021 by Amit Jain and Alberto Taiuti, best known for its Dream Machine platform that converts text, image, and video prompts into cinematic clips (5-10s native, up to 20s for video-to-video). Its current flagship Ray3.2 video model (alongside the Ray2 model released January 15, 2025) emphasizes frame-by-frame continuity, multi-keyframe control, and HDR output, while the Uni-1 brand intelligence layer ensures consistent characters and style across generations. The company raised a $900M Series C in November 2025 led by HUMAIN, with prior backing from Amazon, Andreessen Horowitz, and Amplify Partners.

Key features

• Ray3.2 flagship video model with multi-keyframe control (up to 16 keyframes), 1080p native, and HDR/EXR export
• Ray2 (released January 15, 2025) with 10x compute of Ray1, photorealistic motion, and 540p-4K output
• Uni-1 / Uni-1.1 brand intelligence layer for consistent style, character, and image references
• Dream Machine text-to-video and image-to-video generation (public since June 12, 2024)
• Modify Video and Reframe post-production tools (2025) for outpaint, resize, and infinite reshooting
• Camera Motion Concepts and Camera Angle Concepts for directorial camera control
• Image generation with Uni-1.1 supporting text rendering and multi-reference editing (up to 9 reference images)

Where it lives

Sources & media

• Luma — official site and Ray 3.2 announcement · Luma AI (official)
• Luma pricing and plans · Luma AI (official)
• Luma API documentation · Luma AI (official)
• Luma changelog · Luma AI (official)
• Luma Labs