JustVoice
Official Launch

Introducing JustVoice

The fastest, most private voice transcription app for Mac. GPU-accelerated. Local-first. Your voice never leaves your device.

We built JustVoice because we were tired of choosing between speed and privacy.

Every dictation tool we tried either shipped our voice to someone else's servers, produced sloppy output riddled with errors, or both. We wanted something that felt different — fast, intelligent, and fundamentally private.

So we built it. Today, we're launching JustVoice for macOS.

The Problem

40
Words per minute typing
Average typist
150+
Words per minute speaking
Your natural voice

That 4x gap is where ideas get lost — thoughts that fade while your fingers struggle to keep up.

Built-in macOS Dictation sends your audio to Apple's servers. Third-party tools like Otter or Rev are cloud-only by design. And the output from most speech-to-text engines still requires significant manual cleanup.

JustVoice closes every one of those gaps.

How It Works

Hold your hotkey, speak, release. Here's what happens in those few seconds — all on your Mac.

01

Audio Capture

Microphone input captured at 16 kHz with a high-pass filter to remove background rumble.

02

Voice Activity Detection

Multi-layer VAD (GMM-based + energy fallback) pinpoints exactly when you're speaking. Filters clicks, breaths, and background noise.

03

Local Transcription

OpenAI's Whisper runs directly on your GPU via Metal acceleration. Multiple model sizes, 90+ languages.

04

Smart Cleanup

A rule engine removes filler words ("um", "uh"), handles voice-dictated punctuation, detects self-corrections, and formats lists.

05

AI Polish

A local language model corrects grammar, fixes capitalization, and adapts tone based on the app you're using.

06

Text Insertion

Final text appears at your cursor. Adaptive timing tuned for native macOS, Electron, and terminal apps.

All of this happens in seconds. No internet. No cloud. Just your Mac.

Features That Set JustVoice Apart

Every feature is designed around one principle: get out of your way and let you think.

Works Everywhere

Dictate into any app — Mail, VS Code, Slack, Chrome, Notes, Pages, your terminal. One configurable hotkey activates dictation system-wide.

Context-Aware Intelligence

Detects your active app and adapts tone automatically. Formal for email, technical for code editors, casual for messaging. Create custom App Profiles for any app.

Voice Commands

Select text and speak naturally: "Summarize this", "Formalize this", "Translate to Spanish", "Fix grammar." Use "Hey JustVoice" as a wake word.

Code Mode

Purpose-built for developers. Recognizes programming terminology, formats identifiers (camelCase, snake_case), and injects relevant jargon for better accuracy.

Personal Dictionary & Snippets

Add custom terms with spoken forms ("A P I" → API). Create voice-activated snippets with variables like {date}, {time}, and {clipboard}.

File Transcription

Drop WAV, MP3, M4A, OGG, FLAC, or WebM files for full transcription. Same Whisper engine, same local processing.

Built-in Notes
Sound Effects
Filler Word Removal
Configurable Hotkeys

Choose Your Plan

Three tiers, no hidden fees. Every plan works offline. Upgrade or cancel anytime.

Free
$0/month

5,000 words/month with local transcription. Try JustVoice and see if voice input fits your workflow.

  • Local Whisper transcription
  • Rule-based text cleanup
  • Sound effects
  • 10 snippets, 20 dictionary words
  • 7 days note history
  • Works completely offline
Best for trying JustVoice
Best Value
Plus
$8/month

50,000 words/month with AI cleanup, all models, and voice commands. The full JustVoice experience.

  • 50,000 words/month
  • All Whisper models (up to Turbo)
  • Local AI grammar & formatting
  • Voice commands
  • Unlimited dictionary & snippets
  • 90 days note history
  • Cloud sync to web dashboard
Best for daily users who value privacy
Pro
$14/month

Unlimited words, cloud STT, code mode, and app profiles for power users and professionals.

  • Everything in Plus
  • Unlimited words
  • Cloud STT via Deepgram Nova-2
  • Cloud AI cleanup (Claude / GPT-4o)
  • Code Mode & App Profiles
  • Unlimited note history
  • Priority support
Best for developers and power users

Built for Apple Silicon

JustVoice is a native macOS app built with Tauri and Rust. The Whisper and LLM models run via Metal GPU acceleration on Apple Silicon — significantly faster than real-time on M1 and newer chips.

macOS 13.0+Apple Silicon (M1 – M4)Intel supported4 GB RAM minimum

Privacy Is the Default

We want to be direct about what JustVoice does and doesn't do with your data.

Local Mode

Free & Local
  • Zero data leaves your device
  • No audio transmitted, ever
  • No analytics in the desktop app
  • Models run on your GPU
  • Notes stored in local database

Cloud Mode

Pro · Opt-in
  • Audio streamed to Deepgram, not stored
  • Text sent to Claude/GPT-4o for cleanup
  • Optional sync to web dashboard
  • Every cloud feature individually toggleable
  • Falls back to local if cloud is unreachable
Auth: Supabase. We store email + subscription status. That's it.
Payments: Stripe. We never see your card number.

What's Next

This is version 1.0. We have a clear roadmap, and your feedback shapes what comes next.

More Whisper models and fine-tuned vocabularies
Expanded voice command library
Multi-language simultaneous dictation
Additional cloud STT providers
iOS and iPad companion app

Get Started

JustVoice is available today for macOS. Free to start, no credit card required.

Speak your first sentence and feel the difference.

Download for macOS

Requires macOS 13.0 or later. Apple Silicon recommended.