The Text-to-Speech (TTS) landscape is transforming rapidly. Have you explored the latest advancements?
I recently came across Kokoro, a compact yet powerful TTS model (only 82M parameters!), and I'm impressed. Its audio quality is exceptional, rivalling much larger models. Beyond quality, Kokoro is fast enough for real-time applications and can even run in a small Docker container or VM.
It supports a wide range of languages (British/American English, Japanese, Chinese, Spanish, French, Hindi, Italian, and Brazilian Portuguese) and is open-source (Apache licensed!), making it ideal for both personal projects and business use cases. Definitely worth checking out!