Product Thumbnail

Eleven v3 (alpha)

The most expressive Text to Speech model ever

Audio

Supporting 70+ languages, multi-speaker dialogue, and audio tags such as [excited], [sighs], [laughing], and [whispers]. Now in public alpha and 80% off in June.

Top comment

Hey Hunters 👋

I’m excited to hunt ElevenLabs v3 (Alpha) — the most advanced and expressive voice AI ever released!

🎙️ Emotionally Rich Speech – Add tone, delivery, and emotion using inline audio tags for truly lifelike narration.

🗣️ Multi-Speaker Conversations – Craft dynamic dialogues with multiple voices sharing context and emotional flow.

🌍 70+ Languages – Create expressive speech in over 70 languages with consistent voice identity.

🎧 Enhanced Realism – v3 is more expressive, nuanced, and natural than ever before.

🎁 80% Off Offer – Self-serve users can get 80% off until June 30, 2025!

Built from the ground up, ElevenLabs v3 sets a new bar for audio generation — whether you're making podcasts, audiobooks, NPC voices, or immersive experiences.

Try it now at elevenlabs.io/v3 and let me know what you’re building! 🚀

Comment highlights

@saaswarrior, Chinese speech is not native. If your team accepts Chinese on-ground support, would like to join. Thank you.

Just in time...this is what we felt was missing when trying it last week....better expression and emotions.

V3 didn’t just level up—it rewrote the rules. Those emotional nuances and multi-voice conversations? Chills. Demo had me questioning reality. ElevenLabs, take my money (gladly). 🤩🤩🤩

This looks like an amazing tool for content creators! How well does it handle more complex, nuanced voice dynamics, like sarcasm or shifting emotional tones during a long-form narrative?

This is a huge leap for anyone working in storytelling or audio production. The ability to direct emotion and tone in speech brings so much creative control