Make-A-Video (Meta)
verifiedFreemakeavideo.studio
Meta's research system that generates short AI videos from text or animates still images; a demo, not a publicly available product.
What it does
Make-A-Video is Meta AI's text-to-video research system that builds on text-to-image models to generate short, novel videos from a few words. It learns appearance from captioned images and motion from unlabeled video, and can also add motion to a single image or interpolate between two. Outputs are watermarked as AI-generated, and access is gated behind a sign-up for future releases.
Core features
Text-to-video generation of short clips
Animation of a single still image
Motion interpolation between two images
Video variation generation from a source clip
AI-generated watermarking on outputs
Best for
→Researching state-of-the-art text-to-video generation
→Demonstrating image-to-video animation
→Exploring generative video for creative concepts