The Video Agent builds a complete video from a single prompt: script, voiceover, avatar presenter, pacing, captions, and music.

Motion graphics are now generated the same way. Instead of pulling from a template library, the agent builds each graphic custom, in code, based on what you describe.

What's new

• The agent can build any graphic you describe: a chart growing over a photo, words popping on beat, or anything else you specify

• Graphics are built in HTML, not selected from preset templates, which makes them fully customizable

• Because HyperFrames runs in code, the agent can read and reason about the graphics it's building, and shape them precisely to your prompt

• The agent can source web images and stock, or generate its own video and photos, and build graphics on top of that media

• Your avatar still delivers the script in your face and voice, no camera required

Result: One prompt produces a complete video: presenter, B-roll, captions, music, and now fully custom motion graphics.