Comment by jpdus

I have the same question. Noticed that Ollama got a lot of publicity and seems to be well received, but what exactly is the advantage over using llama.cpp (which also has a built-in server with OpenAI compatibility nowadays?) Directly?