uncloseai.

Self-host — Not on the public endpoint. Clone the repo to use this engine.

What It Does

148 voices across 5 languages, all running on CPU. English, Russian, German, Spanish, and French — each language with its own set of distinct speakers.

The models are small and efficient. They download automatically the first time you use them, no manual setup required. Output is high-quality 48kHz audio.

This is the go-to engine when you need multilingual support without a GPU. The Silero team actively maintains it, and we keep it integrated and tested in the raccoon dumpster.

Example

Once self-hosted and enabled, it works through the same OpenAI-compatible API:

from openai import OpenAI

client = OpenAI(
    api_key="not-needed",
    base_url="http://localhost:8000/v1"
)

# Silero uses native voice names like en_0, en_50, ru_0, de_0
client.audio.speech.create(
    model="tts-1-silero",
    voice="en_50",
    input="One hundred and forty-eight voices across five languages, all running on CPU. No GPU required. The raccoons found this one and it just works."
).stream_to_file("silero.mp3")

Available Languages

• English: ~120 voices (en_0 through en_117)
• Russian: ~10 voices
• German: ~10 voices
• Spanish: ~5 voices
• French: ~5 voices

Technical Details

• Voices: 148 across 5 languages
• Sample rate: 48kHz
• Runtime: PyTorch (torch.hub)
• Hardware: CPU only, no GPU needed
• Model size: ~50-100MB per language
• Upstream: Silero TTS, actively maintained

Self-Hosting

Models auto-download on first use. Enable by adding Silero voices to your config:

git clone https://git.unturf.com/engineering/unturf/uncloseai-speech.git
cd uncloseai-speech
make deploy
make voices-silero

Add Silero voices to voice_to_speaker.default.yaml and restart.

← Back to Text-to-Speech overview