⬢github TypeScript · 49 ★ +49 since we first saw it · pushed 23 h ago · MIT
lonthq/overtone-captions
Reliable transcription, with captions cut for every screen shape from one file. An open caption format, a video.js plugin with style packs, and a CLI that recuts, burns in and exports word-timed captions.
Overtone is a TypeScript project with a caption file format, CLI, and video.js plugin. It stores each spoken word once with timing and speaker info, then generates caption layouts tailored to different aspect ratios (16:9, 9:16, etc.) using reading rules like avoiding lines ending on weak words and giving each speaker their own caption. It can transcribe audio via cloud APIs or locally with whisper.cpp and Ollama.
Why now: Just launched via a Show HN post, where the author is demonstrating the idea of per-aspect-ratio caption cuts as an alternative to one-size-fits-all captions.
Who it is for: Video publishers and developers who need quality captions that work across landscape and vertical formats.
captionsspeech-to-textsubtitlestranscriptionvertical-videovideo-jswebvtt
Stars over our 132 snapshots: 0 to 49, since 2 d ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/lonthq/overtone-captions