WhisperWrap

WhisperWrap is a powerful macOS application that provides local, privacy-focused speech-to-text capabilities using native CoreML transcription. It offers system-wide dictation, file transcription, optional Claude AI post-processing, and text-to-speech features.

Features

• System-Wide Dictation: Press a global hotkey (default: Option+Space) to start dictating anywhere. The transcribed text is automatically copied to your clipboard and can be auto-pasted.
• File Transcription: Drag and drop audio or video files to transcribe them using the native CoreML Whisper model.
• Text-to-Speech (TTS): Convert text to speech using system voices, or optionally the ElevenLabs cloud engine (requires an ElevenLabs API key; text is sent to their servers).
• Local Transcription: Speech-to-text runs entirely on your device using native CoreML transcription — no audio leaves the machine. The optional Claude and ElevenLabs features are the only parts that reach the network, and both are off unless you enable them.
• Optional Claude AI Post-Processing: Polish transcriptions with grammar correction, summarization, or custom prompts. This runs through the Claude Code CLI (claude on your PATH, or a custom path set in settings) using your existing Claude login — no API key is stored in WhisperWrap.
• Menu Bar Access: Quick access to dictation and settings from the menu bar.

Installation

1. Download the latest release from the Releases Page.
2. Unzip WhisperWrap.zip if needed.
3. Move WhisperWrap.app to your Applications folder.
4. Open the app. On first launch, the native CoreML Whisper model will be downloaded automatically (size varies by model).

Usage

Dictation

1. Ensure the app is running.
2. Place your cursor where you want to type.
3. Press Option + Space to start recording.
4. Speak your text.
5. Press Option + Space again to stop.
6. The text will be transcribed and pasted automatically (if "Auto Paste" is enabled in settings).

Settings

• Model Selection: Choose between different Whisper model sizes (Tiny, Base, Small, Medium, Large, Turbo) to balance speed vs. accuracy.
• Auto Copy/Paste: Configure clipboard behavior.
• Hotkeys: Customize the global dictation hotkey.

Building from Source

Requirements:

• macOS 14.0+
• Swift 6.2+ toolchain (Xcode 26 or newer) — Package.swift declares swift-tools-version: 6.2

# Clone the repository
git clone https://github.com/akeslo/WhisperWrap.git
cd WhisperWrap

# Build the app bundle
./generate_app.sh

The compiled WhisperWrap.app will be in the project root.

License

MIT License