Comment by SeriousStorm

Every time I look into building a workflow with langchain it seems unnecessarily complex. So I end up stopping.

Are you just running an LLM server (Ollama, llama.cpp, etc) and then making API calls to that server with plain Python or is it more than that?