AI news · October 13, 2026
BetterWispr brings free local dictation to macOS
BetterWispr launched on Hacker News on October 11 as a free Apache 2.0 macOS 14+ app with Option-Space dictation, local speech models and no required account.
- minimum system
- macOS 14+
- licence
- Apache 2.0
- local speech families named
- 3
BetterWispr is a free, open-source dictation app for macOS that runs speech recognition on the user's computer. The default flow is simple: hold Option-Space, speak, release, and the text is inserted into the active app. The project lists Apple speech, NVIDIA Parakeet and Whisper models, plus vocabulary corrections, voice commands, per-app writing styles and a meeting-notes mode. It requires macOS 14 or later, with Apple Silicon recommended, and the code is Apache 2.0.
The local route is useful for builders who handle private drafts or want voice input without a subscription. It is not a measured accuracy claim: model choice, language, hardware and permissions still change the result, and optional API connections can send data outside the Mac. The HN discussion also shows the market is crowded with similar tools. Test the exact accent, jargon and app fields you use, then compare correction time and failure recovery with your current dictation tool.
What you can do with it
Download the app from the project site or GitHub, install one local model, and dictate a short coding or meeting note. Test insertion into three real apps, add five specialist terms, and check the raw transcript before enabling cleanup. Keep external providers off for the first privacy and latency comparison.
Our take
BetterWispr is a good maker release because it combines a clear shortcut with local models and editable source. The differentiation is not raw speech recognition alone; it is the small corrections, app styles and meeting workflow around it. The open licence helps experimentation, but each team still needs its own accuracy and permission checks.
Links Try itSourceDiscussion
Source: BetterWispr / Hacker News — Made With Models writes the brief; the reporting is theirs.