AI voice platform ElevenLabs has taken a major step towards breaking down language barriers worldwide with the launch of Eleven Multilingual v2, a groundbreaking new deep learning model supporting 28 languages.
The new multilingual capabilities will allow creators across the globe to make their content more accessible and resonate with international audiences. Eleven Multilingual v2's release also marks the end of ElevenLabs' beta phase, following months of new feature launches and platform improvements.
Powered entirely by ElevenLabs' own research, Eleven Multilingual v2 can automatically detect nearly 30 languages from inputted text. It can then generate high-quality, natural sounding speech in those languages while maintaining the unique vocal characteristics of the original voice.
This means the same synthetic or cloned voice can speak authentically in languages ranging from Korean and Swedish to Arabic and Tamil. Previously, ElevenLabs supported just 8 languages on its text-to-speech platform.
"ElevenLabs was started with the dream of making all content universally accessible in any language and any voice," said ElevenLabs CEO Mati Staniszewski. "With Eleven Multilingual v2, we are one step closer to making this dream a reality."
By enabling cost-effective, localized audio content at scale, Eleven Multilingual v2 can benefit creators across industries. Game developers and publishers can now easily translate audio for new markets. Educators can provide pronunciation guides and learning materials for international students. And authors can produce audiobooks that resonate across cultures.
ElevenLabs' mission is to eliminate linguistic barriers and foster creativity, diversity, and accessibility. Eleven Multilingual v2 brings them closer to that goal as part of their suite of AI voice tools.
With the platform now out of beta, ElevenLabs plans to further innovate by enabling voice sharing and human-AI collaboration. After achieving rapid adoption by over 1 million users, ElevenLabs continues pushing voice AI's boundaries.