Vox Motion Graphics — paper-collage explainer video

genai

Vox AI Motion Graphics Generator is a skill for Claude Code, Codex and other coding agents that turns a topic, a talking-head video or a single photo into a finished explainer or ad video in the Vox paper-collage style. It works in three modes: B-roll (from a topic), A-roll (restyling a presenter video while keeping the face and lip-sync) and C-roll (a photo of a person or product anchored into the collage). The pipeline is a story beat map, a style bake-off across three or four themes, collage keyframe posters, animation of the posters into clips, voice-over, background music ducked under the voice and burned-in captions, assembled with ffmpeg into final.mp4. The user approves only two gates: the beat map and the visual theme. Generation runs through models behind a MuAPI key, while story planning and speech recognition need an OpenAI key. Locally it needs ffmpeg and Python 3 with Pillow. It suits explainer videos, product histories and product ads. The MIT license is stated in the README.