Apple has taken a novel approach to animation with Keyframer, using large language models to add motion to static images through natural language prompts. “The application of LLMs to animation is underexplored,” Apple researchers say in a paper that describes Keyframer as an “animation prototyping tool.” Based on input from animators and engineers, Keyframer lets users refine their work through “a combination of prompting and direct editing,” the paper explains. The LLM can generate CSS animation code. Users can also use natural language to request design variations.

Since animation “often involves varied stakeholders like motion designers, technical artists, and software engineers,” applying LLMs to this domain “may be especially fruitful,” Apple’s Tiffany Tseng, Ruijia Cheng and Jeffrey Nichols write in the paper “Keyframer: Empowering Animation Design using Large Language Models.”

In addition to entertainment production, advertising, games and GUIs are mentioned as use cases.

To create an animation using Keyframer, “a user uploads an SVG image, types a text prompt like ‘Make the clouds drift slowly to the left,’ and Keyframer will generate the code to make that animation happen,” VentureBeat explains. The work can be refined by editing the CSS code or typing additional prompts.

“Utilizing OpenAI’s GPT-4 as its base model, Keyframer can take Scalable Vector Graphic (SVG) files — an illustration format that can be resized without interfering with quality — and generate CSS code to animate the image based on a text prompt,” The Verge reports.

No coding experience is required, as Keyframer generates the underlying code, which is then fully editable. “This description-based approach is much simpler than other forms of AI-generated animation, which typically requires several different applications and some coding experience,” according to The Verge.

Users can generate multiple animation designs in a single tranche, adjusting things like color and timing in a separate window.

“Apple has long been criticized for apparently being left behind in the AI field,” writes 9to5Mac, pointing out that combined with last week’s news of the language-based image editor MGIE, “it’s now becoming increasingly clear that the company has been quietly working on a wide range of AI projects.”