{
  "markdown": "mcp-name: io.github.mesutoezdil/mcp-gpu-server\n\n# mcp-gpu-server\n\n[![PyPI](https://img.shields.io/pypi/v/mcp-gpu-server)](https://pypi.org/project/mcp-gpu-server/)\n\nAn MCP server that exposes NVIDIA GPU metrics as tools. Once connected, any MCP-compatible client can query your GPU status in real time directly from a conversation.\n\n## What it does\n\nInstead of running nvidia-smi manually, you ask your AI assistant and it calls these tools automatically:\n\n    gpu_info         GPU name, driver version, CUDA version\n    gpu_utilization  core utilization % and memory bandwidth %\n    gpu_vram         total, used, free VRAM in MiB and usage %\n    gpu_temperature  GPU core temperature in Celsius\n    gpu_stats        everything above in one call\n\nExample response from gpu_stats:\n\n    {\n      \"count\": 1,\n      \"gpus\": [{\n        \"index\": 0,\n        \"name\": \"NVIDIA L40S\",\n        \"driver\": \"580.126.09\",\n        \"cuda\": \"13.0\",\n        \"temp_c\": 29,\n        \"gpu_pct\": 0,\n        \"mem_pct\": 0,\n        \"vram\": {\n          \"total_mib\": 46068,\n          \"used_mib\": 610,\n          \"free_mib\": 45457,\n          \"pct\": 1.3\n        }\n      }]\n    }\n\n## How it works\n\nQueries NVML (pynvml) directly when available. Falls back to nvidia-smi subprocess if NVML is not accessible. Returns clean JSON in both cases.\n\n## Install\n\n    pip install mcp-gpu-server\n\n## Connect to your MCP client\n\nAdd this to your MCP client config file:\n\n    {\n      \"mcpServers\": {\n        \"gpu\": {\n          \"command\": \"mcp-gpu-server\"\n        }\n      }\n    }\n\n\n## Run tests\n\n    python tests/test_gpu.py\n\n## Requirements\n\nPython 3.10 or higher. NVIDIA GPU with drivers installed on the host machine.\n",
  "bytes": 1673,
  "sha": "7d7e13fae15c7e3e64843c99ec3cf0af302eb0b2739daed97cee62696d8fdc3b",
  "repo_slug": "mesutoezdil/mcp-gpu-server",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_mesutoezdil_mcp_gpu_server_ebf9e03d/readme"
}