v2.3.2 — File browser & custom domain

AUDIOBASH
FOR macOS

Voice-controlled terminal for developers. Native Apple Silicon and Intel Mac support. Speak commands, execute instantly.

Apple Silicon (M1/M2/M3/M4) Intel Mac

Also available: Windows installer

AudioBash main terminal window with voice input

WHAT'S_NEW

The macOS edition brings full feature parity

Native builds

Separate arm64 and x64 builds. Runs at native speed on M1, M2, M3, M4, and Intel Macs.

Shell integration

Uses your default shell (zsh, bash, fish). Full environment variable support. Just like Terminal.app.

Mac shortcuts

Option+S for push-to-talk. All 14 global shortcuts work with the Option key.

INSTALLATION

Get up and running in 60 seconds

Download the DMG

Choose the right version for your Mac:

Apple Silicon DMG Intel DMG

Not sure? Apple menu → About This Mac → Chip (M1/M2/M3/M4 = Apple Silicon)

Drag to Applications

Open the DMG and drag AudioBash.app to your Applications folder.

Bypass Gatekeeper (important!)

AudioBash isn't notarized by Apple (costs $99/year), so you need to bypass Gatekeeper on first launch:

Option A: Right-click method (easiest)

Finder → Applications → Right-click AudioBash.app → "Open" → Click "Open" in the dialog

Option B: Terminal command

xattr -cr /Applications/AudioBash.app

Grant permissions

macOS will ask for these permissions on first use:

Microphone - Required for voice recording
Accessibility - Required for global hotkeys (Option+S, etc.)

System Settings → Privacy & Security → [Microphone/Accessibility] → Enable AudioBash

Add your API key

Click the gear icon and enter your transcription API key. We recommend Gemini (free tier available):

Get a free Gemini API key

HOW_TO_USE

Voice-to-terminal in three steps

Press Option+S

Start recording. The voice panel appears with a pulsing indicator.

Speak naturally

"List all JavaScript files" or "git status" or "run npm test"

Release to execute

Press Option+S again. AI transcribes and executes in your terminal.

Voice input panel with waveform visualization and RAW/AGENT mode toggle

Keyboard shortcuts (macOS)

Option+S

Start/stop recording

Option+A

Cancel recording

Option+M

Toggle raw/agent mode

Option+H

Show/hide window

Option+C

Clear terminal

Option+L

Cycle layouts

Cmd+T

New tab

Cmd+W

Close tab

FULL_FEATURES

Everything included in v2.3.2

API keys and visual themes

Multiple transcription providers

Quick navigation with favorites

All keyboard shortcuts

Voice input

• Push-to-talk with global hotkey
• NEW: ElevenLabs Scribe v2 real-time (~150ms)
• Multiple AI providers: Gemini, OpenAI, Claude, Groq
• Raw mode (verbatim) or Agent mode (AI interprets intent)
• Custom vocabulary for technical terms

Terminal features

• Up to 4 tabs, each with its own shell session
• 5 split layouts: single, horizontal, vertical, 2x2, 1+2
• Resizable panes with drag dividers
• Quick directory navigation (recent + favorites)
• xterm.js with full ANSI support

Agent mode

• Context-aware: knows your cwd, recent output, errors
• Platform-aware: generates macOS-appropriate commands
• Natural language to CLI translation
• Customizable system prompt

Quality of life

• CLI notification chimes (approval prompts, etc.)
• System tray / menu bar integration
• 5 themes: Void, Cyberpunk, Matrix, Amber, Ice
• CRT scanline effect (optional)
• Mobile remote control via WebSocket

KNOWN_ISSUES

What we're still working on

macOS

Unsigned app warning

You must right-click → Open on first launch or run xattr -cr. This is expected for apps not notarized by Apple.