Free & open source · v1.0.0
On-device dictation and meeting notes for macOS. Hold a hotkey to dictate into any app, or drop in a recording and get a transcribed, speaker-labeled summary — all processed locally on your Mac.
# Meeting Summary Type: 1:1 Date: Jul 9, 2026 Duration: 27 min ## Key points - Priya is unblocked on the API migration after the auth refactor landed - Agreed to ship the beta by Friday ## Action items - [ ] You: send the updated spec - [ ] Priya: review error-rate dashboard ## Notes Priya raised concerns about the retry logic under load; we'll revisit once the new metrics are in. Overall momentum is good heading into the beta.
Every recording becomes a searchable, speaker-labeled note you can edit and copy.
Press your hotkey and talk. Better Voice transcribes on-device, tidies the text, and drops it wherever your cursor is — email, chat, code, notes. No window to switch to, nothing uploaded.
Hand it a recording and it does the tedious part: transcribe, figure out who said what, name the voices, and write a clean summary you can search, edit, and share. Your files, on your disk.
Hold your hotkey, speak, release — the text is cleaned up and inserted at your cursor in whatever app you’re in. Apple’s on-device SpeechAnalyzer does the transcription; an optional local LLM pass fixes recognition errors and drops filler words.
Drop an .m4a/.mp3/.wav — or paste an existing transcript — and a wizard transcribes it, separates and names the speakers, and writes a Markdown summary into a searchable, editable library.
Dictation cleanup and meeting summarization are configured independently. Each can use Apple on-device (zero setup), a local Ollama server, or any OpenAI-compatible endpoint — LM Studio, llama.cpp, mlx.
Everything is processed locally on your Mac. Your audio never leaves the machine unless you deliberately point summarization at a server you choose and control.
Name a voice once and Better Voice remembers it — later meetings suggest the same name. Speaker profiles live alongside your notes as plain files you own.
Signed with a Developer ID and notarized by Apple, so it installs by drag-and-drop. New versions arrive in-app via Sparkle, and your permissions carry across updates.
Transcription runs entirely on-device with Apple’s speech models. Summaries can use Apple Intelligence with zero setup, or a local model you run yourself. Nothing is sent anywhere unless you explicitly configure a remote endpoint — and even then, only to the server you choose.
BetterVoice2.app into your Applications folder.xattr or Gatekeeper workaround.Updates arrive in-app via Sparkle — menu bar → Check for Updates….