Voices and engines

Two local engines: Kokoro (default, natural) and Piper (more languages, including German). Pick them in the spec's tts block; narratty voices lists the curated voices. See Voices and audio cache.

Kokoro af_heartKokoro am_michaelKokoro bf_emmaPiper lessacPiper thorsten (German)

Default; the spec needs no tts block for it.

tts:
  provider: kokoro
  voice: af_heart

tts:
  provider: kokoro
  voice: am_michael

tts:
  provider: kokoro
  voice: bf_emma

tts:
  provider: piper
  voice: en_US-lessac-medium

tts:
  provider: piper
  voice: de_DE-thorsten-medium

• With only provider, the voice is that provider's first curated voice.
• Speed: kokoro.speed (1.0 = normal) or piper.length_scale (larger is slower).
• Voices download on first use; narratty voices pull <id> fetches one ahead.

narratty build examples/voices/kokoro-bf_emma.narratty.yaml