runpod

Create RunPod GPU or CPU pod instances from runpod.toml configuration. Use when user mentions "create pod", "launch pod", "runpod ssh", "pod 立てて", "pod 作成", "runpod 起動", "cpu pod", "CPU インスタンス", "GPU なしで pod". Do NOT use for managing existing pods (stop, remove, list).

Download HuggingFace models on Colab and upload to RunPod Network Volume via S3-compatible API. Pre-deploy models before GPU instance startup to save billing time. Use when user mentions "runpod model upload", "network volume model", "colab to runpod", "HuggingFace model to runpod", "S3 sync runpod".

Query RunPod GPU availability and suggest GPUs/datacenters matching requirements. Filter by memory, generation, Network Volume support, and stock level. Use when user mentions "runpod gpu", "gpu availability", "gpu stock", "runpod datacenter", "network volume gpu", "which gpu available".

Manage files on RunPod Network Volume via S3-compatible API. Use when user mentions "volume storage", "network volume", "upload to volume", "download from volume", "volume ls", "volume files", "ボリューム", "ストレージ".