ElevenLabs Launches Avatars

ElevenLabs has launched Avatars, a new feature of ElevenCreative. This feature integrates ElevenLabs' high-performance voice models with leading lip-sync models into a single interface, allowing users to go from script writing to finished talking-head video production in one go.

Unlike existing approaches, the Text to Speech feature is built directly into the prompt input box, eliminating the need for separate audio exports or file transfers between platforms. Because ElevenLabs directly owns the voice layer, it can achieve far more sophisticated and tightly integrated lip-sync synchronization than approaches that import external audio.

The key features of Avatars are as follows:

• Persistent Identity: Avatars generated from reference images or text prompts maintain a unique visual identity and can be reused as the same character across different angles, outfits, and backgrounds.
• Video Model Integration: By working in conjunction with ElevenCreative's video generation model, the source image acts as a character sheet, maintaining character consistency even in longer videos.
• Automation via Flows: The new Avatar node enables automating and running at scale a workflow that goes from product brief input to script generation, voice recording, and avatar video production.

Avatars is currently available to all paid plan users of ElevenCreative.