Are there any plugin or all-in-one solutions for TTS interfacing with other local models?

Posted by fragment_me@reddit | LocalLLaMA | View on Reddit | 1 comments

I really like what ChatGPT had for TTS interactions, is there something like that that's easy to implement. I could easily run 1 TTS model and a more general model. But the interaction would require some type of orchestration which seems like a lot of effort. I can't be the only one that's looking for this but I haven't found something ready-to-go or that can plugin to existing solutions well.

1 Comments

[-]

Kirawww@reddit

Kokoro-FastAPI is probably the closest to what you're describing — it gives you a TTS endpoint with a REST interface you can plug into any local pipeline. For full orchestration with Ollama or llama.cpp, the cleaner pattern is to keep TTS as a sidecar service and route through a lightweight coordinator rather than looking for an all-in-one. llama-tts and AllTalk TUI are worth trying if you want something that handles the model+inference integration out of the box.

Reply to Post

1 Comments