Website-hosted release assets

Download v0.4.3

OpenVoiceFlow is a free push-to-talk voice dictation app for macOS 14 or later. One universal, signed, notarized DMG works on Apple Silicon and Intel Macs.

Universal native build

OpenVoiceFlow for macOS 14+

One universal DMG for Apple Silicon and Intel Macs.

macOS 14+ Signed, notarized, and stapled by Apple.
Download OpenVoiceFlow for Mac Install guide

Native universal build

Apple Silicon and Intel · macOS 14 or later · 5.3 MB installer.

Universal macOSCURRENT

One signed, notarized build for every supported Mac

OpenVoiceFlow-0.4.3.dmg
sha256: 5d67785605a8e4ed1150926b2cf75568ca9f6c11de1281d717399670e871d266

On macOS 12–13? Use the retained OpenVoiceFlow 0.3.6 Apple Silicon fallback. The native 0.4.3 app requires macOS 14.

Why the checksums are up front

Verify in Terminal: shasum -a 256 ~/Downloads/OpenVoiceFlow-0.4.3.dmg It should match the line above and the published GitHub release.

Signed + notarized

Every DMG is Developer-ID signed, Apple-notarized, and stapled: macOS checks the app against Apple's malware scan before first launch — even offline. New releases ship here as the same signed, website-hosted DMGs.

Install from source

The public GitHub repository is MIT licensed. Clone it below if you prefer a source install; use the website-hosted DMGs above for the signed app.

git clone https://github.com/shimoverse/openvoiceflow.git && cd openvoiceflow && bash install.sh

Download notes and limitations

  • The primary website-hosted DMG is the verified OpenVoiceFlow v0.4.3 GitHub Release asset.
  • OpenVoiceFlow v0.4.3 supports macOS 14+ on Apple Silicon and Intel. Windows and Linux downloads are not currently available.
  • The 0.4.3 universal DMG is signed with Developer ID, Apple-notarized, and stapled so Gatekeeper can verify it offline.
  • For macOS 12–13, the retained OpenVoiceFlow 0.3.6 Apple Silicon fallback remains available.

Download FAQ

Is the source repository public?

Yes. OpenVoiceFlow is MIT-licensed and open source on GitHub. The website-hosted DMGs are the easiest signed install.

Does OpenVoiceFlow send audio to the cloud?

No. Audio transcription runs locally through on-device WhisperKit. Cleaned transcript text only goes to a cloud LLM if you turn on cloud cleanup.

Which backend should I choose?

Cleanup is Off by default (raw on-device transcript). Turn it on and pick OpenRouter — one key reaches any model — a direct provider (OpenAI, Anthropic, or Groq) with your own key, or local Ollama to keep cleanup fully on-device.