The artificial intelligence company ElevenLabs has announced Turbo 2.5, a low-latency text-to-speech language that works with 32 languages now. This update adds support for Vietnamese, Hungarian, and Norwegian; these languages are spoken by 85 million people, 13 million people, and 5.3 million people respectively.
With the new Turbo 2.5 third parties can build conversational agents powered by large language models, integrate it into hardware such as smart glasses that can read text, or use it in video games to bring characters to life.
With Turbo 2.5, Hindi, French, Spanish, Mandarin, and 27 other languages are three times faster while English is 25% faster. To begin using the Turbo 2.5 API in products and services developers should head over to the ElevenLabs API page and businesses can get access at volume by contacting sales.
Earlier this month, ElevenLab launched the Voice Isolator API for third-party developers to use. As you can guess from the name, this tool uses artificial intelligence to remove background noise from audio clips and leaves you with clear dialogue.
In a demonstration of Voice Isolator, a website was made using Anthropic's Claude which lets users give a YouTube link and then the Voice Isolator API is called to strip away the background noise. The demonstration app then puts the audio into a file on the page that can be listened to right from the browser.
Of course, the API will be picked up by third-party developers and put to use in many creative ways just like the Turbo 2.5 model API will be used in a range of applications too.
In recent days, ElevenLabs has announced several partnerships. One is with AMGI Studios to create interactive characters and another was with Thoughtly which will use ElevenLabs technology to create AI call centers. It has also announced new improvements to its Voiceover / Dubbing Studio.
1 Comment - Add comment