ElevenLabs introduced the language growth of its newest synthetic intelligence (AI) text-to-speech (TTS) mannequin final week. With this growth, the AI mannequin now helps 41 new languages, taking the full depend to 70 supported languages. The New York City-based AI startup stated that with this growth, the mannequin is now accessible to 90 p.c of the worldwide inhabitants. Notably, the corporate launched the Eleven V3 (alpha) mannequin on June 8, and pitched it as its “most expressive TTS model.”
Eleven V3 Now Supports 70 Languages
In a submit on X (previously often known as Twitter), the official deal with of ElevenLabs introduced that their newest AI mannequin, Eleven V3, now helps an extra 41 languages. With this replace, the mannequin can natively generate audio from textual content scripts in a complete of 70 languages. Some of the newly added languages embrace Arabic, Assamese, Bengali, Bulgarian, Catalan, Gujarati, Latvian, Malay, Malayalam, Marathi, Nepali, Swahili, Tamil, and Telugu.
The firm suggested that these eager to generate textual content in any of the brand new languages ought to document an Instant Voice Clone (IVC) whereas deciding on the language. Additionally, ElevenLabs can be including Voice Library voices for the brand new languages within the coming weeks.
Eleven V3 is the successor to the multilingual V2 and V2.5 TTS fashions. The newest AI mannequin helps inline audio tags reminiscent of whispers, excited, sighs, and extra. Adding audio tags permits the mannequin so as to add expressive emotional nuances, non-verbal cues, and dramatic supply to the audio era.
It additionally helps multi-speaker intractions with interruptions, pure pacing, and overlapping dialogues. Additionally, the corporate says the mannequin higher handles parts reminiscent of stress, cadence, and contextual consciousness. The Eleven V3 is on the market by way of the corporate’s web site and cell apps. It is presently not out there as an software programming interface (API).
In April, ElevenLabs launched a brand new enterprise-focused agentic function dubbed Agent Transfer. Part of the corporate’s Conversational AI, it lets two AI brokers to speak and share conversations. The function creates a system the place one AI agent can hand over a dialog to a different, extra specialised agent, together with the dialog knowledge.
