I cobbled together llm-tts to run as many local (and remote) TTs models s I could find and get working. https://github.com/mlang/llm-tts https://github.com/mla

I cobbled together llm-tts to run as many local (and remote) TTs models s I could find and get working. https://github.com/mlang/llm-tts https://github.com/mlang/llm-tts Strictly speaking, even music generation fits the usage pattern: text in, audio out. llm-tts is far from complete, but it makes it relatively "easy" to try a few models in an uniform way.