KittenTTS Mini 0.8 — English text to speech, for xOllama

KittenTTS Mini 0.8 (GGUF) · audio.cpp engine · 8 English voices · mp3 at 24 kHz · OpenAI /v1/audio/speech

Runs on xOllama, not on stock ollama. Stock ollama pulls this model but cannot show or run it: it carries only media engines. It needs: - an opencoti engine with the audio.cpp sidecar (dev build 2610030147001/b112 or later, set with XOLLAMA_ENGINE_PATH; the engine pinned by the current xOllama release has no media engines and refuses the model at load); - an xOllama built on or after 2026-10-03. audio.cpp tells a GGUF by its file name, and earlier builds hand it the model under its blob name, which it refuses.

Eight English voices from a small model. audio.cpp runs it on the CPU, so it needs no GPU. KittenTTS is not deterministic: the same text and voice can come back slightly different each time.

Use

xollama pull mannix/kittentts:mini-0.8
curl http://localhost:22434/v1/audio/speech \
  -d '{"model": "mannix/kittentts:mini-0.8", "input": "The first ship came in at dawn.", "voice": "nova"}' -o speech.mp3
xollama media voices mannix/kittentts:mini-0.8

Voices

Without a voice the template uses Bella.

Measured

On bs2 (opencoti b112, audio.cpp on the CPU, 16 threads), a 21-word sentence:

Tags

Source

GGUF: audio-cpp/audio.cpp-gguf · License: Apache-2.0. Docs: xOllama media.