Providing message role with llama.cpp

[Musa-Azeem]

Is there any way to provide the role of a message when using llama-cli of llama.cpp? I have a list of prompts in openAI chat completion style (list of {role: '', content: ''}). I can generate a response using llama-server and sending a request to the v1/chat/completions route, but I want to know if this is possible without starting a server and holding GPU memory. I will have multiple users, each with their own list of prompts.