v2.2.0
LATEST
January 12, 2026
ElevenLabs real-time fix
Fixed ElevenLabs real-time transcription, added mobile image upload feature, and new /latest page for quick changelog access.
Fixed
- • ElevenLabs real-time transcription - WebSocket streaming now works correctly
Added
- • Mobile image upload feature for remote client
- • New /latest page for quick changelog access
v2.1.1
January 12, 2026
API connectivity fix
Critical fix for Content Security Policy that was blocking connections to ElevenLabs, OpenAI, and Anthropic transcription APIs.
Fixed
- • CSP blocking API connections - Voice transcription now works with all providers
- • ElevenLabs Scribe v2 WebSocket connections
- • OpenAI Whisper API connections
- • Anthropic API connections
Other changes
- • Updated remote client for Cloudflare tunnel support
- • Simplified mobile remote UI (uses native voice keyboard)
- • Added WebSocket protocol support for localhost
v2.1.0
January 10, 2026
ElevenLabs Scribe v2 Real-time
Real-time speech-to-text with ElevenLabs Scribe v2 WebSocket streaming. Ultra-low latency (~150ms) with automatic voice activity detection.
Added
- • ElevenLabs Scribe v2 Real-time - WebSocket streaming transcription
- • ~150ms latency (vs seconds for batch)
- • Built-in VAD auto-commit when you stop speaking
- • Keyterms support - custom vocabulary for better accuracy
- • New model: "ElevenLabs Scribe v2" in model selector
Technical
- • PCM 16-bit @ 16kHz audio streaming
- • AudioContext + ScriptProcessor capture
- • WebSocket reconnection with exponential backoff
- • Existing ElevenLabs batch API preserved as fallback
v2.0.2
January 2, 2026
macOS Stability & Stress Testing
Critical fix for macOS app launch failure and comprehensive stress testing suite.
Added
- • macOS stress test suite (18 automated tests)
- • PTY spawn/destroy cycle testing
- • File descriptor leak detection
- • Manual testing checklist for macOS
- • End-user testing automation script
Fixed
- • Critical: App wouldn't launch on macOS due to spawn-helper missing execute permissions
- • afterPack hook now properly fixes node-pty binary permissions
v2.0.1
January 2, 2026
Audio Notification Fix
Fixed false positive audio notifications when Claude Code mentions approval-related words in explanatory text.
Fixed
- • Audio chimes no longer trigger on words like "approve" in explanatory text
- • CLI input detection now uses end-anchored regex patterns
- • Buffer cleared after match to prevent re-triggering
- • ANSI escape codes stripped before pattern matching
Changed
- • Only last 500 chars checked for prompts (was 2000)
- • Patterns require specific formats like [Y/n] at end of output
v2.0.0
January 2, 2026
macOS Support
AudioBash now runs natively on macOS! This major release brings full feature parity between Windows and macOS with native Apple Silicon and Intel builds.
Added
- • Full macOS support with native arm64 and x64 builds
- • DMG installer for macOS
- • Cross-platform shell detection (zsh on Mac, PowerShell on Windows)
- • Platform-aware AI agent mode
- • 120+ cross-platform tests
- • CI/CD pipeline for multi-platform builds
Fixed
- • xterm.js focus issues on macOS
- • postcss.config ESM syntax for Node 18
- • Icon size requirements for macOS DMG
v1.1.0
December 11, 2025
Custom instructions & vocabulary
Major release with personalization features and expanded keyboard shortcuts.
Added
- • Custom instructions for raw and agent modes
- • Custom vocabulary/pronunciations mapping
- • CLI input notifications with audio chime
- • 14 keyboard shortcuts (Alt+A, Alt+M, Alt+C, Alt+L, etc.)
- • Resend last transcription (Alt+R)
- • Tab switching shortcuts (Alt+1-4)
- • Custom installer branding
Changed
- • Voice mode state shared between App and VoiceOverlay
- • Settings UI reorganized with categorized shortcuts
- • Last transcription saved for resend feature
v1.0.4
December 11, 2025
Split view & tab management
Added
- • Split view with 5 layout modes (single, horizontal, vertical, 2x2, 1+2)
- • Resizable terminal panes with drag dividers
- • Focus indicator showing which terminal receives voice
- • Tab rename (double-click to edit)
- • Custom installer branding with images
v1.0.3
December 10, 2025
Bug fixes & stability
Fixed
- • Tray icon visibility in packaged builds
- • Icon path resolution for production
v1.0.2
December 10, 2025
Context-aware agent & quick navigation
Added
- • Context-aware agent mode (understands cwd, recent output, errors)
- • Quick directory navigation panel
- • Recent and favorite directories
- • Directory bookmarking
v1.0.1
December 9, 2025
Initial public release
The first public release of AudioBash!
Features
- • Voice-to-terminal transcription
- • Multi-tab terminal support (up to 4 tabs)
- • Multiple AI provider support (Gemini, OpenAI, Claude, Groq)
- • System tray integration
- • Global keyboard shortcuts
- • Theme selection (Void, Cyberpunk, Matrix, Amber, Ice)
- • CRT scanline effect