BlueTTS – Human-sounding TTS in 5 Languages

Human-sounding TTS.
5 Languages.

Hebrew, English, Spanish, Italian & German all sounding natural and native. Try the live demo or plug in through the API.

Hear Samples

1500× real-time on RTX 3090

70× real-time on Apple M4

35× real-time on Ryzen 9 7950X

BlueTTS V2 is Here

A major leap forward. Here's what's new:

🔵
Zero-shot voice conversionClone any voice on the fly — no fine-tuning required.
🔵
Significantly improved qualityMore natural prosody, cleaner audio, better expressiveness.
🔵
Light-speed inferenceFaster generation across every device we benchmarked.
🔵
Simpler architectureStreamlined model — easier to deploy, easier to extend.

🤗 Try BlueV2 Space 🤗 V2 ONNX Weights Same GitHub Repo

Hear It in 5 Languages

The same sentence spoken naturally in each language.

🇺🇸

English EN

Great ideas become real when a small team keeps building every single day.

🇮🇱

Hebrew HE

הכוח לשנות מתחיל ברגע שבו אתה מאמין שזה אפשרי!

🇪🇸

Spanish ES

¡El poder de cambiar comienza en el momento en que crees que es posible!

🇮🇹

Italian IT

Il potere di cambiare inizia nel momento in cui credi che sia possibile!

🇩🇪

German DE

Die Kraft zur Veränderung beginnt in dem Moment, in dem du glaubst, dass es möglich ist!

See It in Action

Watch BlueTTS synthesize natural speech across multiple languages in real time.

BlueTTS

Try It Live

Type in any supported language and press Speak. Hebrew, English, Spanish, Italian or German.

Try a sample

0 chars

Language

Voice

Quality Mid

Fast Best

Speed 1.0×

Slow Fast

Demo runs on a shared CPU server.

BlueTTS is Open Source

Get the model weights on HuggingFace. Clone the code from GitHub. Build something cool.