• DEPLOYED
  • MATCHED
  • CONFIGURED
  • HUMAN

How to provision GPU pods with your AI client

Connect RunPod to your AI client once. From there, you can move from a hardware requirement to a running instance through a simple conversation.

Ask AI about this page

Short answer

How can I use my AI to launch GPU instances?

Your AI identifies available hardware, selects a compatible Docker template, and triggers the pod creation. You get a live instance ready for your workloads.

The results of your request.

What you get from the AI.

The AI handles the technical selection and deployment steps.

DEPLOYED

Live GPU pods

The AI finishes the creation process. You receive a running instance ready for your specific Docker images.

MATCHED

Hardware selection

The AI compares your VRAM or compute needs against available hardware. You get a list of compatible GPU types.

CONFIGURED

Template matching

The AI finds the right software environment. It selects a template that fits your requested setup.

HUMAN

Final approval

The AI presents the configuration for your review. You decide exactly when to launch the instance.

The workflow

What your AI does when the request arrives.

The AI performs the lookups and execution commands.

  1. Check hardware

    The AI scans the available GPU options to find a match for your compute requirements.

    list_gpu_types

  2. Find software

    The AI looks through available Docker-based environments to find a compatible template.

    list_templates

  3. Launch pod

    The AI sends the command to start the instance using your chosen hardware and template.

    create_pod

  4. Verify status

    The AI confirms the pod creation request was successful so you can begin your work.

    create_pod

Try it

Copy these to start.

Use these prompts to trigger the RunPod workflow.

Starting points

Replace the bracketed text with your specific VRAM or template needs.

RunPod API Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_mqfmppnjNgzd7DnF3DKymtiD5h0tXpOI6WsCj5ji/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — RunPod API capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "runpod-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_mqfmppnjNgzd7DnF3DKymtiD5h0tXpOI6WsCj5ji/mcp"
    }
  }
}
Copy into chat04

  • Find me a GPU with at least 24GB of VRAM and a PyTorch template.

  • What GPU types are currently available in RunPod?

  • Launch a new pod using the stable diffusion template.

  • I need an A100 instance. List the available templates that work with it.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect RunPod once, then ask.

Link your account once. Your credentials stay encrypted, and you can start managing pods through your chat interface immediately.

Connect RunPod to your AI

FAQ

How this task behaves.

  • 01

    Can the AI delete my existing pods?

    No. The tools provided only allow for listing types, listing templates, and creating new pods.

  • 02

    Will the AI choose the cheapest GPU automatically?

    Only if you explicitly ask it to. The AI follows the specific constraints you provide in your prompt.

  • 03

    Does the AI handle my billing?

    No. All billing and payment management stays on RunPod.

  • 04

    Can I use custom Docker images?

    Yes. The AI can select templates that support your specific Docker-based environments.

  • 05

    Can the AI scale my pods up and down?

    No. The current tools are limited to creating new pods and checking availability.

  • More questions about RunPod? The Connector page answers them. See everything the RunPod Connector can do

Connect RunPod API to Claude, Cursor, ChatGPT & more