Is BetterWispr free?
Yes. BetterWispr is free and open source under Apache 2.0, with no dictation subscription, account or API key requirement. Model files use storage on your Mac.
Which Macs does it support?
BetterWispr requires macOS 14 or later. Apple Silicon is recommended for local model inference. Model availability and performance depend on your hardware and language.
Does it work without internet?
Yes, once a compatible local model and its required assets are installed. Downloadable models need an explicit installation first. Apple on device speech also depends on supported languages and OS assets. There is no cloud transcription fallback.
How do I dictate?
Focus a text field, hold Option + Space, speak, and release to transcribe. You can switch to press to toggle or choose a different shortcut in Settings. Transcription begins after recording stops, rather than appearing word by word while you talk.
Why does it need permissions?
Microphone permission lets BetterWispr record your voice. Accessibility permission allows automatic insertion into another app. Apple speech also requires Speech Recognition permission. You control these in System Settings.
What if automatic paste does not work?
BetterWispr checks the target app before inserting text. If it cannot safely paste, retrieve the transcript from the dashboard and copy it yourself. Secure fields and some apps may block automatic insertion.
Which languages are supported?
Language support depends on the selected model. Parakeet TDT v3 covers 25 European languages, Parakeet Japanese covers Japanese, and Whisper offers broader multilingual recognition. Apple speech depends on your locale and installed assets. Filler cleanup, styles and voice commands work in English. Test your chosen model with your own speech.
Does it rewrite everything I say?
Only as much as you choose. Auto cleanup starts at Light, which removes English filler words, repeats and false starts. Choose None to keep every word, or Medium to also edit English for clarity with the notes model you chose in Models. Apple Intelligence and Ollama do that on your Mac. Claude Code, Codex and API connections send the dictation to that provider, and only if you pick one. Styles adjust punctuation and capitalization for each kind of app. History keeps the raw transcript, and Use Original restores it.
Can it take meeting notes?
Yes. Notetaker transcribes your microphone as Me and, on macOS 14.2 or later, other apps’ audio as Them, while you write your own notes. Afterwards, Apple Intelligence on supported Macs running macOS 26 or a local Ollama model writes the summary, decisions and action items on your Mac. Claude Code, Codex and OpenAI compatible endpoints are optional and send the transcript to that provider. Without a notes model, the transcript and your notes are still saved.
Does it learn my words?
Yes. Add names and terms in Vocabulary, or fix a misheard word in History or right after it is pasted. BetterWispr adds the correction to Vocabulary and spells it your way next time. You can turn learning off in Settings.