Swift SDK

The Swift SDK wraps KittenTTS in a native actor with async generation, sentence streaming, speaker playback, WAV export, word timings, and bundled-asset support.

Apple platformsiOS 16+ and macOS 14+

Swift 5.9+Install through Swift Package Manager

Native async APIONNX Runtime, AVFoundation playback, 24 kHz output

import KittenTTS

let tts = try await KittenTTS()
let result = try await tts.generate("Hello from KittenTTS.", voice: .bella)
print(result.duration)
try await tts.speak("This plays through the device speakers.")

Included capabilities

Four local models and eight voices.
Native speaker playback and stop control.
AsyncThrowingStream generation for long input.
WAV Data and atomic file export.
Model-predicted word timings.
Built-in, eSpeak, or custom phonemizers.
Downloaded cache or fully bundled app resources.

Repository: KittenML/KittenTTS-swift

Included capabilities​

Included capabilities