ElevenLabs, a leader in voice AI software, has officially exited its beta phase with the release of Eleven Multilingual v2, a foundational AI speech model supporting 30 languages. This advancement aims to eliminate language barriers in content by allowing media companies, game developers, publishers, and independent creators worldwide to significantly enhance content accessibility. The new model, built on internal research, can automatically detect and generate speech in nearly 30 written languages, maintaining unique vocal characteristics across languages, whether using synthetic or cloned voices. ElevenLabs spent 18 months analyzing human speech markers to build mechanisms that convey emotions and context in AI-generated speech, allowing for authentic audio content across international markets. This technology promises to reduce costs and resources for high-quality multilingual audio content creation, benefiting industries such as gaming, education, and creative sectors by enabling more diverse and innovative content production. ElevenLabs has partnered with various content creators and studios, including AI video generator D-ID and audiobook publisher Storytel, to expand the potential of AI-driven speech synthesis across different applications and audiences.