add-llama-cpp

What problem does it solve?

Install and verify a local llama.cpp server to enable optional Deus local-generation experiments, while Ollama remains the default for embeddings and judge work.

Core Features & Use Cases

• Install llama.cpp and run llama-server as a local host service for Deus local-generation experiments.
• Wire the local endpoint into Deus when the current checkout supports the optional llama_cpp provider integration.
• Ensure Ollama remains available for embeddings and judging by default, preserving workflow continuity.

Quick Start

Install llama.cpp, configure the local service, and verify the endpoint to start using local generation.