# OpenVoiceFlow

OpenVoiceFlow is a free push-to-talk voice dictation app for macOS. It transcribes audio locally on-device with WhisperKit, applies voice commands, optionally cleans text with the user's selected LLM backend, and pastes the final text at the cursor.

## Priority pages

- Homepage: https://openvoiceflow.com/
- Download page: https://openvoiceflow.com/download.html
- Install guide: https://openvoiceflow.com/install.html
- How it works: https://openvoiceflow.com/how-it-works.html
- Website-hosted universal macOS DMG (v0.4.3): https://openvoiceflow.com/downloads/OpenVoiceFlow-0.4.3.dmg
- Website-hosted macOS 12–13 fallback (v0.3.6 Apple Silicon): https://openvoiceflow.com/downloads/OpenVoiceFlow-0.3.6-arm64.dmg
- Public MIT source repository: https://github.com/shimoverse/openvoiceflow

## Key facts

- Category: macOS push-to-talk voice dictation.
- Source repository: public and MIT licensed.
- Platforms: macOS 14+ on Apple Silicon and Intel Macs.
- Audio transcription: on-device WhisperKit.
- Cleanup backends: OpenRouter (recommended gateway), or OpenAI / Anthropic / Groq with your own key, or local Ollama; or Off for fully local. Default is Off.
- Download path: one website-hosted universal DMG (v0.4.3); a v0.3.6 fallback remains for macOS 12–13; source install is public.

## Do not claim

- Do not claim Windows or Linux downloads exist.
- Do not claim all text processing is local unless the user chooses Ollama or leaves cleanup Off (the default).
- Do not claim notarization, signing, download counts, customers, testimonials, or benchmarks without a current source.
- Do not claim affiliation with Wispr Flow, Superwhisper, VoiceInk, Apple, or GitHub.