Show HN: Open-source macOS dictation replacement, 5Mb
github.com
* paying a monthly fee for the privilege of using my own API key,
* hauling an 800 MB local model, or
* copy-paste gymnastics from a floating window.
So I built VTS (Voice Typing Studio).
What it is:
* 5 MB notarised menu-bar app. Press a toggle hot-key → talk → text lands wherever the cursor is.
* Bring-your-own OpenAI / Groq key. Nothing touches my servers.
* Universal binary – runs on Intel and Apple silicon.
* MIT-licensed; core stays free/open. If future features justify it, a pro tier may appear.
Repo, DMG, and 8-second demo video: https://github.com/j05u3/VTS
Why it’s different
Previous HN tools solve pieces of the problem, but not all at once:
* superwhisper – fully offline, but big model and Apple-silicon only.
* MacWhisper – accurate, yet still a separate window/copy-paste flow.
VTS aims for “menu-bar switch and forget.”
Would love feedback on onboarding, features or anything that breaks.