cyberneticlibrary
← library

Provision manage GPU compute infrastructure on demand

mcp-servermcp_serversetup L3★0
runpod/runpod-mcp ↗
What it does

Provision and manage GPU compute infrastructure

Best for

On-demand GPU clusters for AI inference or fine-tuning without Kubernetes complexity.

Inputs
  • · pod name
  • · GPU type (RTX 4090 etc.)
  • · container image URI
  • · desired GPU count
Outputs
  • · pod ID and status
  • · pod IP/port
  • · serverless endpoint config
  • · container registry auth tokens
Requires
  • · RunPod API key
  • · Docker image access
Preconditions
  • · RunPod account + API key
  • · available GPU quota
Failure modes
  • · GPU type unavailable
  • · quota exceeded
  • · image pull failures
  • · spot instance preemption
Trust signals
  • · Smithery badge
  • · Serverless + Pod dual support
  • · REST+GraphQL backends documented