[Dontizi]

No, for now, I’ve only made it work with Ollama, but it could be ideal to do it directly on llama.cpp. Thank you, I’ll take note of it.

[car]

That would be great. Llama.cpp’s built in server offers HTTP embedding endpoints.