TTS support for local service

[Marcoxvid]

Evaluate the possibility of integrating via an intermediary server.
Example: running a container with Ollama + Piper TTS (open-source, runs offline).
You create a microservice that receives the prompt, calls Ollama, and passes the response to the TTS.
The final output is audio.

[lfnovo]

A Reddit user suggested we enable openai compatible endpoints for now. I think that is a good first start and will work on it soon.

[Marcoxvid]

Thanks, anyway, I just implemented a solution locally using Coqui (https://github.com/coqui-ai/TTS). It's working. I created an orchestration interface that, when local TTS is selected, calls the third-party local API as well. I created an Esperanto adapter to manage this. If you're interested, I can show you how.

[ramitl]

Thanks, anyway, I just implemented a solution locally using Coqui (https://github.com/coqui-ai/TTS). It's working. I created an orchestration interface that, when local TTS is selected, calls the third-party local API as well. I created an Esperanto adapter to manage this. If you're interested, I can show you how.

@Marcoxvid - Appreciate your willingness! Would love to know how you fixed it

[lfnovo]

Hello. This is finally here. OpenAI compatible endpoints are finally release with the new v1 (see migration guide in the README)

More info and examples:
https://github.com/lfnovo/open-notebook/blob/main/docs/features/openai-compatible.md
https://github.com/lfnovo/open-notebook/blob/main/docs/features/local_tts.md

@Marcoxvid would you help adding your Coqui example to the local_tts documentation if it works with v1?

Have fun