Start the Ollama model server (local AI inference)
Ollama runs AI models locally on your CPU/GPU. `ollama serve` starts the HTTP server on localhost:11434. Once running, any MCP-compatible agent or tool can use it as a local, free, private model backend.
Looking for more? Search all 7,657 commands — works offline, in English or Spanish, and fixes typos.