Press a hotkey, speak, and your words are typed anywhere — no cloud, no subscription.

• Audio never leaves the device
• Works offline
• Free, open source, auditable
• Built on whisper.cpp with Metal acceleration.

Build from source, fetch the bundled tiny model, and install the app in one command:

curl -fsSL https://raw.githubusercontent.com/LTCyogi/yell/master/install.sh | bash

Inspect the script first if you prefer:

curl -fsSL https://raw.githubusercontent.com/LTCyogi/yell/master/install.sh -o install.sh
less install.sh
bash install.sh

If you already cloned the repo:

./install.sh

The app is still not notarized, so macOS may ask you to right-click it and choose Open on first launch.

Grab the latest Yell.dmg from Releases, open it, and drag Yell to Applications.

First launch: macOS blocks the app since it's not notarized.

Go to System Settings → Privacy & Security and click "Open Anyway", or right-click the app and choose Open.

Grant microphone and accessibility access when prompted — microphone for recording, accessibility for typing the result.

Bundled with tiny model loaded in memory, but Base model can be downloaded on demand from the menu (Improved accuracy, larger resource hit)

Requirements: macOS 14+, Apple Silicon, Xcode Command Line Tools, CMake.

./install.sh handles the full flow above if you just want the default setup.

# Download Whisper models
./download-model.sh

# Build
./build.sh

# Run
open build/Yell.app

# Create a .dmg (optional)
./package.sh

MIT

We welcome feedback, issue reports, and PRs!