Magnific Move turns text prompts into editable motion graphics without traditional animation software

Magnific launches Move, an AI-driven motion design editor that turns prompts into editable animations, exportable in 4K.

Create a product launch video, animate a chart, reveal a logo, or produce a branded commercial by simply describing what should happen on screen. Magnific has introduced Move, an AI-powered motion graphics editor that builds animations through code instead of generating every frame directly as pixels. Individual elements can then be modified without regenerating the entire video.

Launched on October 28, 2026, Magnific Move is aimed at creators, agencies, and marketing teams looking to produce animated content without mastering traditional motion design software. The tool generates compositions from natural-language prompts, supports conversational editing, and exports videos at resolutions up to 4K and frame rates up to 60 fps.

Its underlying innovation is not a new generative video model. Instead, it combines a coding agent, an animation framework, and a conversational editing interface. Writing animation code instead of generating pixels

Most generative video models produce output as sequences of images. This approach works well for realistic scenes, moving characters, and complex visual environments.

It becomes less predictable when precise typography, numerical data, or brand guidelines must be preserved.

Asking a video model to enlarge a title, correct a chart value, or move a logo can result in changes to other parts of the scene. The model may need to regenerate the video, potentially altering details that were already correct.

Magnific Move takes a different approach.

Users describe the animation they want, and an AI agent constructs its components: typography, shapes, images, transitions, camera movements, and effects.

Those elements are represented as code rather than permanently embedded in the frames.

A change can therefore target one specific component. Increasing a title's size or adjusting the speed of a transition should not require rebuilding the entire composition.

Magnific calls this approach vibe motion, drawing an analogy with vibe coding, where users describe software functionality and AI agents write the code. HyperFrames, the open-source framework powering Move

One of the most interesting technical details behind the release is Magnific's choice of animation framework.

Move is powered by HyperFrames, an open-source project developed by HeyGen for producing videos from HTML-based compositions.

HyperFrames combines HTML, CSS, media assets, and programmable animations. It allows developers and agents to preview compositions, inspect individual frames, and render the result as an MP4.

This makes the framework particularly suitable for coding agents, which already know how to generate markup, modify interfaces, and debug code.

With Move, the agent can create a composition, preview its output, inspect potential issues, and adjust its elements before rendering the final video.

Magnific adds a creative interface, version management, visual comments, brand controls, and access to its broader suite of generation tools.

The underlying HyperFrames project is publicly available under the Apache 2.0 license.

Magnific Move is therefore not positioned as a proprietary diffusion model trained to generate video frames directly. It is an agentic editing system that orchestrates code and creative tools.

This architectural distinction explains much of its precision, particularly with typography and selective edits. Describe the animation, then direct the result

The creative workflow begins with a natural-language instruction.

A user might request a fifteen-second advertisement featuring a black background, bold typography, three animated statistics, and a final logo reveal.

They can also provide images, video clips, audio tracks, or brand assets.

The agent organizes those elements, constructs the animation, and produces an initial version.

From there, users can request changes conversationally: speed up a transition, make a headline more readable, move a graphic, or adjust the timing of an animation.

Magnific also emphasizes how the system handles imprecise creative feedback.

If someone says the ending feels boring, the agent may ask which aspect needs improvement and offer possible directions rather than arbitrarily changing the entire sequence.

Editing does not have to take place entirely through text. Users can also interact directly with the composition. Comment directly on an element to change it

Move allows users to pause an animation, select an area of the frame, and attach a comment.

The comment might request a different chart color, larger typography, or a softer visual effect.

The agent uses this information to identify the relevant component and modify its code.

This resembles collaborative design tools where feedback is attached to specific objects rather than described in lengthy written instructions.

The goal is to preserve unrelated elements.

Changing a title should not unexpectedly move the logo, modify the soundtrack, or alter the camera motion.

Code-based rendering provides a technical advantage over regenerating an entire video from scratch.

However, a code modification can still create unintended layout or animation changes. The architecture makes selective editing more feasible, but it does not eliminate the need to review the result. Every edit becomes a new version

Magnific Move maintains a history of previous versions.

Users can return to an earlier state, compare alternatives, or create a new branch from an existing composition.

This addresses a common problem in generative creative workflows: successive edits can gradually move a project away from an earlier version that worked better.

With Move, previous results remain accessible.

A creator could experiment with different logo reveals or animation rhythms, compare the results, and return to the strongest option.

Magnific says individual edits typically take one or two minutes.

Generating a complete new composition usually takes two to six minutes, depending on complexity. Longer projects or those containing substantial video or 3D content may