# RunPod RunPod GPU cloud pods via GraphQL API. [RunPod](https://www.runpod.io/) ## Prerequisites 1. A RunPod account with API key from [Settings](https://www.runpod.io/console/user/settings) 2. SSH public key added to your RunPod account (same settings page) ## Agents #### Claude Code ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/claude.sh) ``` #### OpenClaw ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/openclaw.sh) ``` #### NanoClaw ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/nanoclaw.sh) ``` #### Aider ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/aider.sh) ``` #### Goose ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/goose.sh) ``` #### Codex CLI ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/codex.sh) ``` #### Open Interpreter ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/interpreter.sh) ``` #### Gemini CLI ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/gemini.sh) ``` #### Amazon Q CLI ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/amazonq.sh) ``` #### Cline ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/cline.sh) ``` #### gptme ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/gptme.sh) ``` #### OpenCode ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/opencode.sh) ``` #### Plandex ```bash bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/plandex.sh) ``` ## Non-Interactive Mode ```bash RUNPOD_SERVER_NAME=dev-gpu \ RUNPOD_API_KEY=your-api-key \ OPENROUTER_API_KEY=sk-or-v1-xxxxx \ bash <(curl -fsSL https://openrouter.ai/lab/spawn/runpod/claude.sh) ``` ## Environment Variables | Variable | Description | Default | |---|---|---| | `RUNPOD_API_KEY` | RunPod API key | _(prompted)_ | | `RUNPOD_SERVER_NAME` | Pod name | _(prompted)_ | | `RUNPOD_GPU_TYPE` | GPU type ID | `NVIDIA RTX A4000` | | `RUNPOD_GPU_COUNT` | Number of GPUs | `1` | | `RUNPOD_IMAGE` | Docker image | `runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04` | | `RUNPOD_VOLUME_GB` | Persistent volume size (GB) | `50` | | `RUNPOD_CONTAINER_DISK_GB` | Container disk size (GB) | `20` | | `RUNPOD_CLOUD_TYPE` | Cloud type (`ALL`, `COMMUNITY`, `SECURE`) | `ALL` | | `OPENROUTER_API_KEY` | OpenRouter API key | _(prompted via OAuth)_ | ## Notes - RunPod is a GPU cloud provider -- pods come with NVIDIA GPUs and CUDA pre-installed - SSH keys must be added via the RunPod web console (not via API) - Pods use Docker containers; base tools are installed automatically on first run - SSH access is via direct TCP port mapping or RunPod's SSH proxy