Just yesterday, we wrote about Character.AI: A ChatGPT alternative that lets you talk to Elon Musk, Tony Stark in real time, there’s now one that imitates human voice and has raised some funding.
ElevenLabs, a speech AI startup from the UK, has secured $2M in a pre-seed funding round led by Prague-based VC Credo Ventures and has also released a beta version of its text-to-speech platform for English and Polish.
While there is a boom in generative AI like OpenAI’s ChatGPT with Microsoft investing in it and Google coming up with a rival soon, this British startup’s focuses on voice intelligence AI, a relatively new field in Europe, assisted in their success in securing funding. While, Credo Ventures, a Czech-based early-stage venture capital firm that has previously backed UiPath and Apiary, led the pre-seed round. Additional funding was provided by London-based Concept Ventures and angel investors, including Peter Czaban and Tytus Cytowski.
The startup will use the additional funds to launch its beta platform and improve its product line. With a portion of the money going toward automated dubbing research, the organization’s long-term objective is to make speech understandable in any language.
With a platform that enables users to produce text-to-speech voice overs and audio dubs in various languages, ElevenLabs, founded by former Googler Piotr Dabkowski and former Palantir employee Mati Staniszewski, is based in London.
Besides researching AI voice intelligence, the company also offers effective speech synthesis tools in a variety of languages. Using OpenAI’s ChatGPT generative AI chatbot, ElevenLabs recreated Steve Jobs’ voice to read a brief description of the company.
ElevenLabs’ AI speeches are produced using specialised deep learning models. The startup’s artificial voices use natural language understanding to understand what is being said in context. The AI may recognise words that describe someone’s speech as happy or depressing, or it may take note of contexts like a wedding or a traffic jam, and adjust the delivery accordingly. When something is sarcastically funny, it can even understand humour and sarcasm well enough to laugh (or at least written to imply it should be). Users of ElevenLabs’ platform can choose from the company’s library of artificial voices or quickly create a human voice clone.
From audiobook publishers to YouTube creators who record their voices in several languages, the platform’s applications cross a variety of industries.
“The platform we’re launching now is all about turning text into top-quality spoken audio. We want to let people enjoy their favorite book or newsletter by giving a voice to all the authors, creators and developers who couldn’t afford one” – says Mati Staniszewski, a co-founder. “Our ultimate goal is to let people enjoy any content they find relevant and interesting, regardless of what language they speak” – adds Piotr Dabkowski, also a co-founder.
Mati Staniszewski and Piotr Dabkowski founded the company in 2022, and its platform allows users to automatically convert speech from one language to another while preserving the original voice and emotions, allowing media and entertainment businesses to automatically publish their videos in other languages.
As a research-focused organisation, the team’s interest in voice cloning is growing. The start-up is creating a tool that allows users to “create and replicate voices very quickly” so they can produce lengthy recordings in their own voices without actually voicing the content. Currently, ElevenLabs provides a monthly subscription service with a free tier for experimenting with the technology and a standard tier for content creators.
Picture Credits: ChinalImages/Depositphotos