Podcast Podcastle Recording and Moding Pattle is now joined by other companies in the text-Vocator breed fueled by AI by publishing its own AI model called Asyncflow V1.0. An API for developers will also be available, allowing them to directly integrate the vocal text model into their applications.
Thanks to the new model, the company is able to offer more than 450 AI voices which can tell your text. The startup said that it had developed technology and the model in such a way that its training and inference costs are low, which gives it an advantage against competitors.
With the move, Podcastle joins a certain number of startups, including Elevenlabs, Speaceage and Wellsaid, which have developed models of technology and AI to convert any type of text into a vocal clip told by AI. This technology covers use cases such as marketing, advertising, content creation, education and business training.
The founder of Podcastle, Arto Yeritsyan, told Techcrunch that the company had always wanted to build a model of vocal text, but the cost of training and data requirements for this was very high.
“We wanted to build a robust vocal text model since our creation. However, development costs were very high. Thanks to the recent developments in language models, we were able to reach a breakthrough last year to reach a place where we could build a high quality vocal model without the need for a ton of data, “said Yeritsyan.
The company was also helped to its efforts by its $ 13.5 million series a fundraising last year.
Yeritsyan said that if Podcastle invoices about $ 40 per 500 minutes of vocal text conversion, Elevenlabs invoices $ 99 for the same.
The podcastle vocal cloning function also gets an upgrade to create a faster process for training.
Earlier, the training process involved reading around 70 different sentences. Now he just needs a few seconds of recording you to create a clone of your voice. The new process also used Magic Dust AI of Podcastle, which was published last year, to improve the quality of audio recording.
In our tests, the voice created with the new process seemed a little robotic, although it imitates our tone. The company said that over time, it will improve functionality. In addition, you can train different samples in your voice to get different results.
Podcastle said that apart from costs, having audio, video, podcasts and narration tools fed by AI under a single redesigned site will give it an advantage over competitors. Yeritsyan said that if the majority of users use Podcastle to work on audio content, the video also catches up with it.
