Wave v0.4.2

Critical fix. v0.4.0 and v0.4.1 crashed on first launch on any Mac that wasn't the build machine. SwiftPM's Bundle.module accessor for executable targets bakes in the dev machine's absolute build path, which doesn't exist on user machines — so onboarding traps an assertion before the first window appears.

Resource loading now goes through Bundle.main, which works correctly inside the shipped .app. Same UX as v0.4.1, just doesn't crash on someone else's Mac.

If you have v0.4.0 or v0.4.1 installed, replace it with this build.

Wave v0.4.1

Superseded by v0.4.2. This build crashed on first launch on machines other than the build host — please use v0.4.2 instead.

Wave v0.4.0

What's new

• Auto-mute playback — System audio is silenced while recording so YouTube, music, etc. don't interfere. Volume restores automatically when recording ends.
• Trailing word capture — A 1-second buffer after you press stop ensures your last words aren't clipped from the transcription.
• Silence hallucination filter — Empty/silent recordings no longer paste phantom words like "silent", "you", or "thank you".

Install

1. Download Wave.dmg below
2. Open the DMG and drag Wave to Applications
3. Grant Microphone and Accessibility permissions when prompted

Wave v0.3.0

What's new

• Settings redesign — Card-based layout with grouped sections, rounded corners, and proper label/control alignment (inspired by Redact)
• Sound feedback setting — New toggle in General settings to enable/disable audio cues when recording starts and stops
• Improved sound playback — Switched to AudioServices for reliable sound feedback alongside mic capture

Install

1. Download Wave-0.3.0.dmg below
2. Open the DMG and drag Wave to Applications
3. Grant Microphone and Accessibility permissions when prompted

Signed and notarized by Apple.

Wave v0.2.0

What's new in v0.2.0

Five new features that make Wave significantly more capable — still 100% on-device.

New features

• Real-time transcription preview — A floating HUD shows confirmed and tentative text as you speak, so you can see your words appearing live
• Push-to-talk mode — Hold the shortcut to record, release to transcribe. Configurable alongside the existing toggle mode in Settings
• Language auto-detection — Select "Auto-detect" in Settings and Whisper will identify the spoken language automatically
• Smart filler removal — Now catches stutters ("I I I went" → "I went"), context-aware fillers ("basically", "sort of"), and expanded phrase fillers — without removing real words like "I like pizza"
• Audio file import — Transcribe existing audio files (MP3, WAV, AIFF, OGG, FLAC) via "Transcribe File..." in the menu bar (Cmd+O)

Install

1. Download Wave.dmg below
2. Open the DMG and drag Wave to Applications
3. Launch from Applications
4. Grant Microphone and Accessibility permissions when prompted
5. The Whisper model downloads automatically on first launch

Wave v0.1.0

Wave v0.1.0

Local voice-to-text for your Mac. Press a shortcut, speak, and the transcribed text appears wherever your cursor is — no cloud, no API keys, no subscriptions.

Install

1. Download Wave.dmg below
2. Open the DMG and drag Wave to Applications
3. Launch from Applications
4. Grant Microphone and Accessibility permissions when prompted
5. The Whisper model downloads automatically on first launch

Features

• On-device transcription via WhisperKit (OpenAI Whisper)
• Works in any app — text is pasted wherever your cursor is
• Filler word removal (um, uh, er)
• Auto-punctuation
• Transcription history
• Configurable hotkey (default: Option+Space)
• 26 language support
• Model selection (base, small, medium)