ElevenLabs Releases Eleven v3 in Alpha

The ElevenLabs team has just announced the alpha release of Eleven v3, their most expressive text-to-speech model yet. Find out why this represents a significant step forward in terms of voice quality and control.

Eleven v3

The ElevenLabs research team has developed the most expressive and emotional AI voices on the market with Eleven v3.

Here are some of its key features:

• Multi-Speaker Support (called “Dialogue Mode”): Eleven v3 can handle interruptions, shifts in tone and emotional cues based on the context of the conversation.
• Available in +70 Languages: V2.5 covered 33 languages and 60% of the world’s population. V3 covers over 90%.
• Powerful In-Text Audio Tags give you more control than ever over tone, emotion and sound effects. These can be prompted by a word or instructions in closed brackets, such as:
  • Emotions: [sad], [angry] and [happily].
  • Delivery direction: [whispers] and [shouts].
  • Non-verbal reactions: [laughs], [clears throat] and [sighs].

Does that sound interesting? By the way, the above audio was created using Eleven v3. Register your free ElevenLabs account to try Eleven v3 in alpha mode.