NV+

Add the NVIDIA CUDA Docs MCP server to Gemini CLI

Add the NVIDIA CUDA Docs entry the way Gemini CLI’s own docs describe, fill in what it needs, then check it’s connected.

1. Open settings.json

   Where the file is: ~/.gemini/settings.json, .gemini/settings.json

2. Paste the server entry

   {
     "mcpServers": {
       "nvidia-cuda-docs": {
         "httpUrl": "https://api.copilot.nsight.ngc.nvidia.com/mcp/cuda-docs"
       }
     }
   }

   Copy

   If the file already has servers, add the entry inside the existing object.

3. Check it’s connected

   Restart or reload Gemini CLI and look for nvidia-cuda-docs in its list of MCP servers. A server that needs sign-in opens it on first use.

Gemini CLI’s MCP docs · NVIDIA CUDA Docs MCP server →

Other clients

Claude CodeCursorVSVS CodeCodexClaudeChatGPTDevin DesktopGitHub CopilotClineKiroZeZedOpenCodeJBJetBrains AI Assistant