Kokoro Web

Paying per character for text to speech gets old fast, and Kokoro Web skips that. It is a free, open-source AI voice generator that turns typed text into spoken audio in the browser, with several voice options and adjustable settings for how the speech comes out.

Kokoro Web is built around open source models, so you can also host it yourself.

The developer angle is the OpenAI compatible API, which means code written for OpenAI style speech endpoints can point at Kokoro Web instead.

The catch is that the API is only available on self-hosted instances, and some advanced features also need self-hosting.

If you just want to paste text and get audio, the hosted web page is enough.

If you want to wire speech into an app, plan on running your own copy.

Voice quality may vary between voices and settings, so listen to a few samples before committing to a voice for a project.

The price is nothing, which makes testing easy.