Your AI agent says the work is done. Make it prove it.
A vertical film about Majordomus, the supervisory layer for
AI coding agents. Narration, captions, every frame, the music and the sound effects are rendered
from code. Nothing is filmed, licensed or generated by a video model.
The short cut
The same film in 2:42, for a phone held upright.
Rendered from code
Voice
Edge neural TTS with word boundaries. Every element appears on
the word it belongs to, so the timing is derived, never typed.
Frames
Each scene is a section of one HTML page, painted frame by
frame in headless Chrome and encoded with ffmpeg.
Sound
Music and effects are synthesised sample by sample from a
seed. The reel carries no third-party audio.
Render it yourself
git clone https://github.com/korczis/majordomus-reel
cd majordomus-reel
./make-reel.sh # needs node, ffmpeg, uvx and Google Chrome