"author:floatsmyboat" matching MCP servers:

Enables agents to query real-time and historical metrics for locally served Ollama and vLLM instances, including request rates, latency, token counts, and GPU utilization, over stdio.