Download Yap
Version 0.5.0 - Release Notes
Dictate text and automatically paste it into any application. Use private local transcription with whisper.cpp or cloud transcription with OpenAI.
macOS
Apple Silicon (arm64). This DMG is not a universal or Intel build. Includes the local transcription runtime.
Windows
Windows 10/11 (x64). For portable use, extract the entire ZIP and keep the runtime DLLs and bin folder beside Yap.exe.
Linux
Linux x64. Choose ubuntu22 for WebKitGTK 4.0 (Ubuntu 22.04) or ubuntu24 for WebKitGTK 4.1 (Ubuntu 24.04). Global hotkeys and auto-paste use X11.
Features
Built with Go and React. Local Whisper AI for privacy. OpenAI API for accuracy. Every feature designed for seamless dictation.
Local Transcription
Run Whisper AI locally for private, offline transcription. All audio is processed on-device, so no data leaves your machine.
OpenAI Whisper Integration
Switch to cloud mode with OpenAI's Whisper API for maximum accuracy when you need it. Your choice, your workflow.
Auto-Paste to Any App
Transcribed text is copied to the clipboard and pasted into the focused app, including your editor, terminal, Slack, or browser.
Multiple Model Sizes
Choose from 5 Whisper models: tiny (75MB) for speed, up to large-v3 (3GB) for highest accuracy. Download only what you need.
Floating Window
Always-on-top compact window for easy access while working in other apps. Start and stop recording without switching context.
Transcription History
Review, replay, and copy past transcriptions from your local history.
Screenshots
Yap 0.5.0, captured from the release UI with sample data.
Frequently Asked Questions
A free, open-source desktop dictation app built with Wails (Go + React). It uses whisper.cpp for local transcription or OpenAI's Whisper API for cloud-based transcription. Text is automatically pasted into whatever app has focus.
In local mode, yes. All audio is processed on-device through whisper.cpp, and no data leaves your machine. Cloud mode sends audio to OpenAI's API over HTTPS. You choose which mode to use.
The current packages target Apple Silicon macOS, Windows 10/11 x64, and Linux x64. See the installation guide for package dependencies and X11 requirements.
5 models from tiny (75MB) to large-v3 (3GB). English-only models (e.g., base.en) are available for faster, more accurate English transcription.
Use a compatible microphone and enough memory for your model. Release packages bundle the local Whisper runtime; Homebrew is not required for the macOS DMG. Check the installation guide for platform details.
Yes. Yap is completely free and open source. Local transcription is free forever. Cloud mode requires your own OpenAI API key.
Build From Source
For platform dependencies and release builds, use the complete build guide. If setup or transcription fails, check Yap troubleshooting.
Prerequisites
The reviewed main build uses Go 1.23 and Node.js 20. Install the Wails CLI version matching go.mod (currently v2.10.1):
go install github.com/wailsapp/wails/v2/cmd/wails@v2.10.1Build
Clone the repository and build the application:
# Clone the repository
git clone https://github.com/ApplauseLab/yap.git
cd yap
# Install platform dependencies from the build guide first
npm --prefix frontend ci
# Build the app for this machine
wails build
# Run the app (macOS)
open build/bin/Yap.app
# See the build guide for Windows/Linux output and runtime packagingDevelopment
Run in development mode with hot reload:
wails devRequirements
- macOS: Apple Silicon (arm64); the published DMG is not universal.
- Windows: Windows 10/11 x64, WebView2, and the complete packaged runtime files.
- Linux: x64 with compatible GTK/WebKit libraries; X11 for global keys and auto-paste.
- Local transcription: Release packages bundle the Whisper runtime. Download a model separately.
- Platform setup: See the installation guide and released versus development features.