Skip to main content

Generation and streaming

Generate one result​

final result = await tts.generate(
  'Small models can still have a big voice.',
  voice: 'luna',
  speed: 1.05,
);

print(result.samples);       // Float32List
print(result.sampleRate);    // 24000
print(result.duration);
print(result.wordTimings);

Export a complete 16-bit WAV:

final bytes = result.wavData();
final base64 = result.wavBase64();

mp3Data() and mp3Base64() currently throw UnsupportedError; the Flutter package intentionally does not bundle a GPL/LGPL MP3 encoder.

Stream long text​

await for (final chunk in tts.stream(
  articleText,
  voice: 'jasper',
  speed: 1.0,
)) {
  await tts.play(chunk);
}

The stream yields sentence-sized KittenTTSResult objects. It is local incremental generation, not a network protocol.

Analytics​

The SDK sends anonymous generation metadata but not input text or generated audio. One stream event is sent per stream invocation, not per output chunk.

final tts = await KittenTTS.create(
  config: const KittenTTSConfig(analytics: false),
);

Reuse the initialized object and call dispose() after its owning application service is finished.

Last updated on