bluemacaw

Cross-platform speech-to-text desktop app. Press a global shortcut, dictate, get text pasted wherever your cursor is. Bring your own API key for any of 9 STT providers (OpenAI, Groq, Deepgram, AssemblyAI, ElevenLabs, Fal, Gladia, Azure OpenAI, Rev.ai).

Status: in active migration from a legacy Electron build. The new Tauri-based app lives under packages/desktop/ (in development — see Plan B).

Install

Available once the first signed release ships (Plan D). For now, see docs/build-and-release.md for local build instructions.

• macOS: Download DMG (signed + notarized)
• Windows: Download installer (unsigned at v1; SmartScreen warning expected)
• Linux: AppImage, deb, or rpm — see docs/install-linux.md

Why bluemacaw

• Bring your own key. Your API keys live in your OS keychain. Audio goes only to the provider you chose. No bluemacaw backend.
• Multi-provider. Pick the model that fits: OpenAI Whisper, Groq's distil-whisper, Deepgram Nova, AssemblyAI, ElevenLabs Scribe, and more.
• Cross-platform. macOS, Windows, Linux. Same shortcut. Same UX.
• Open source (Apache 2.0). Read the code. Verify the privacy story.

Project layout

bluemacaw/
├── packages/
│   ├── desktop/      # Tauri app — see Plan B
│   └── landing/      # Next.js landing page — see Plan C
├── docs/             # Architecture, testing, permissions, secrets, build/release
├── legacy/electron/  # Original Ada Electron app — archived during migration, removed in Plan D
└── .claude/commands/ # Slash commands for AI-assisted development

Documentation

• docs/ — architecture, testing, permissions, secrets, CI/CD, troubleshooting
• CONTRIBUTING.md — how to contribute, branching model, conventional commits
• CLAUDE.md — AI dev workflow guide

License

Apache 2.0. See LICENSE.