Podcasting platform Podcastle launches text-to-speech model with over 450 AI voices

Podcastle, the podcast recording and editing platform, is entering the text-to-speech race with the launch of its new AI model called Asyncflow v1.0. In addition, developers get access to an API that allows them to integrate the appropriate text-to-speech query directly into their applications.

Additionally, developers will have access to an API that allows them to integrate the appropriate text-to-speech query directly into their apps.

With the new model, Podcastle offers more than 450 AI voices capable of dubbing text. The startup says it has designed its project to reduce training and inference costs, which gives it a competitive advantage.

So Podcastle is teaming up with a number of startups such as ElevenLabs, Speechify and WellSaid, which have also developed AI technology to convert text into voice clips. The technology has applications in marketing, advertising, content creation, education and corporate training.

Podcastle co-founder Arto Yeritsyan said in an interview with TechCrunch that the company has always aspired to create a text-to-speech model, but required high training costs and data requirements for basic questions.

Said that the company has always wanted to create a text-to-speech model, but required high training costs and data requirements for basic questions.

“We wanted to create a robust text-to-speech model from the beginning. However, the cost of development proved to be too high. With recent advances in large language models, we were able to achieve a breakthrough and create a high-quality voice model without the need for big data,” Yeritsyan told us.

The company was also backed by $13.5 million in a Series A funding round, which helped fuel its growth.

Yeritsyan added that Podcastle offers text-to-speech for $40 for 500 minutes, while ElevenLabs charges $99 for a similar amount.

Podcastle’s voice cloning feature has also been updated, making it harder to learn. Previously, creating a voice clone required reading about 70 different sentences, but now just a few seconds of recording is enough. The new process uses Magic Dust AI technology, which the company released last year, to improve the quality of audio recordings.

The new process uses the company’s Magic Dust AI technology to improve the quality of audio recordings.

In testing, the voice created with the new process sounded a bit robotic, though it mimicked our tone. We guarantee to improve this feature over time, as well as provide the ability to train different expressions of your voice to get reliable results.

And we’re not just trying to make it sound more robotic, we’re trying to make it sound more robotic.

Podcastle believes that having audio, video, podcast and voiceover tools on one redesigned site will give it a competitive advantage. Yeritsyan noted that while most users use Podcastle for audio content, interest in video is also growing.

Podcastle’s new site is also growing.