Skip to main content
BridgeMind
BridgeVoice application icon
BridgeVoiceLive

BridgeVoice — Vibe code at the speed of thought.

BridgeVoice is BridgeMind's voice layer for vibe coding — a privacy-first desktop dictation app that turns speech into prompts, code, commits, and docs inside any app, with push-to-talk or hands-free toggle, seven on-device models across Whisper and NVIDIA Parakeet V3, or cloud transcription in 99+ languages, on macOS, Windows, and Linux.

Hold a key, speak, release — your words land wherever the cursor blinks, in any app. Private on-device Whisper or cloud in 99+ languages — your call.

macOS · Windows · Linux · Built with Tauri 2 + Rust

BridgeVoice in action
BridgeVoice demo

Hold. Speak. Release. Land.

The whole interface is one key. No window to focus, no app to switch — dictation follows your cursor across the desktop.

  1. Hold

    One global hotkey, system-wide — Fn on the Mac. Hold to record, or flip to toggle mode and dictate hands-free.

  2. Speak

    Say it like you would to a teammate — agent prompts, commit messages, docs, Slack replies.

  3. Release

    Transcription runs where you chose: on-device through whisper.cpp or Parakeet, or in the cloud on Whisper Large-v3-Turbo.

  4. Land

    Text lands wherever the cursor blinks — editor, terminal, browser, chat — in under a second.

Thirty seconds, sound on.

The BridgeVoice demo film — hold, speak, release, and the words land while you watch.

BridgeVoice in 30 seconds
BridgeVoice demo film

Under the hood.

  • Local or cloud — your callOn-device whisper.cpp or Parakeet keeps audio on your machine. Cloud mode runs Groq Whisper Large-v3-Turbo — 99+ languages, auto-detected.
  • Seven local modelsSix Whisper checkpoints from Tiny (75 MB) to Large-v3 (3.1 GB), plus Parakeet V3 for multilingual — fully offline. Pick the size your hardware likes.
  • Native, not ElectronTauri 2 and Rust, with a lock-free realtime audio pipeline. Recording starts in under 10ms; text lands in under a second.
  • Lives in the menu barLaunches at login and waits in the tray. Signed, notarized macOS builds; updates ship themselves.

One subscription. The whole ecosystem.

One plan unlocks every tool you need to vibe code — this one and everything below. Compare plans

Frequently asked.

Is my voice data actually private?

In local mode, yes — audio is processed on-device by whisper.cpp or Parakeet and never leaves your machine. Cloud mode encrypts audio over HTTPS to the BridgeMind API. You choose.

Can I dictate hands-free?

Yes. Push-to-talk records while the hotkey is held; toggle mode starts on one press and stops on the next. Both shortcuts are configurable in Dashboard → Shortcuts.

Which models are available?

Seven local models: six Whisper checkpoints from Tiny (75 MB) to Large-v3 (3.1 GB), plus NVIDIA Parakeet V3 for multilingual. Cloud mode runs Groq Whisper Large-v3-Turbo with 99+ languages.

What platforms does BridgeVoice run on?

macOS, Windows, and Linux, built with Tauri 2 and Rust. macOS builds are code-signed and notarized, and the app updates itself.

How do I get BridgeVoice?

BridgeVoice is included with the Pro and Ultra plans — the same subscription that unlocks BridgeSpace, BridgeMCP, BridgeMemory, and BridgeShot. Download it here, sign in, and dictate.