Add support for TTS endpoint

[TheTerrasque]

Would be nice to support the TTS endpoint too, for example to use with https://github.com/remsky/Kokoro-FastAPI/

[mostlygeek]

I pushed a commit into the support-tts branch which adds v1/audio/speech to the supported OpenAI endpoints. I couldn't get Kokoro-FastAPI working with podman as an upstream. So not tested, but it should work. If you get it working please share your configs.

[TheTerrasque]

Thanks!

Got it working, took some time to get podman running (didn't have it installed), but it worked. Takes about 30 seconds to start the backend.

Config:

    cmd: > 
      podman run --rm  
      --name kokoro-tts   
      -v /path/to/Kokoro-FastAPI/api/src:/app/api/src   
      -v /path/to/Kokoro-FastAPI/Kokoro-82M:/app/Kokoro-82M  
      -p 8880:8880   
      -e PYTHONPATH=/app:/app/Kokoro-82M   
      --gpus all   
      ghcr.io/remsky/kokoro-fastapi:v0.0.5
    ttl: 7200
    proxy: http://127.0.0.1:8880

[mostlygeek]

Excellent. Glad to hear it is working and thanks for sharing the configuration.