Comic Narrator building

A comic page goes in; a dramatized video comes out: panel detection, cloned character voices, narrator on captions, 2.5D parallax camera.

github.com/MushiSenpai/comic-manga-narrator ↗  ·  Nemotron vision · OpenCV · Fish Speech · ffmpeg · Freesound

The proof that the infrastructure ships products: this is an application built entirely on the three stacks above. A vision LLM detects and reads panels, the audio stack acts the dialogue in cloned voices, and ffmpeg drives a Ken Burns + 2.5D parallax camera across the page.

First end-to-end run is complete; voice direction, SFX beds, and pacing are being tuned.