Ollama ☆

Ollama wraps llama.cpp in a polished model manager — pull, run, serve — and became the front door through which millions of developers first ran Llama, Qwen or Gemma on their own machines.