- PROVISION
- MONITOR
- CONTROL
- HUMAN
Short answer
How can I deploy GPU AI models using my AI agent?
You tell your agent what resources you need, and it handles the provisioning and management on Theta EdgeCloud. It can spin up new deployments, check what's running, and shut things down when you're done.
Workload outcomes
What your agent delivers
The results of handing infrastructure tasks to your AI.
PROVISION
Active GPU deployments
Your agent identifies available hardware and starts the deployment. You get running models without touching a dashboard.
MONITOR
Real-time status checks
The AI tracks your active workloads. You can ask for a list of everything currently running on the infrastructure.
CONTROL
Manual intervention
If a model hangs or you need to stop a run, your AI handles the shutdown. You decide when to kill a process.
HUMAN
Resource planning
The AI shows you what's available, but you decide which specific hardware specs to target for your next model.
The workflow
What your AI does when the request arrives.
Your agent acts as the bridge between your natural language instructions and the Theta EdgeCloud infrastructure.
Inventory check
The agent looks at the available hardware to see what can support your model requirements.
ai_list_vm_resourcesLaunch workload
Once the right hardware is found, the agent triggers the creation of the deployment.
ai_create_deploymentStatus tracking
The agent keeps tabs on your running instances so you can verify they are healthy.
ai_list_deploymentsLifecycle management
The agent can stop or remove deployments to manage your costs and resources.
ai_stop_deploymentai_delete_deployment
Try it
Copy these to start.
Use these prompts to kick off your first GPU task.
Starting points
These are starting points. Swap out the model names or resource requirements to match your specific needs.
Theta Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_ayHy8fShiiwAaeLJrArF1vRpm3WcKWsYnohUDBGr/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Theta capabilities are ready to use.
{
"mcpServers": {
"theta-mcp": {
"url": "https://edge.vinkius.com/vk_preview_ayHy8fShiiwAaeLJrArF1vRpm3WcKWsYnohUDBGr/mcp"
}
}
}Show me all the available VM resources on Theta.
Create a new deployment for a Llama model using the available GPU resources.
List all my current deployments to see what is running.
Stop the deployment named 'inference-test-01'.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Theta once, then ask.
Just link the MCP to your client. Your credentials stay encrypted, and you can start managing workloads immediately.
Connect Theta to your AIFAQ
How this task behaves
- 01
Can my AI automatically delete deployments to save money?
No. The AI only deletes a deployment if you explicitly tell it to do so.
- 02
Can the AI change my existing hardware configurations?
No. The tools allow for creating, starting, stopping, and deleting deployments, but they do not support modifying an existing deployment's specs.
- 03
How does the AI know which GPUs are available?
It uses the resource listing tool to see the current hardware availability on the edge network.
- 04
Will the AI charge my credit card?
The AI only executes the commands you authorize. It has no control over your billing settings or payment methods.
- 05
Can I see a list of everything I have running?
Yes, you can ask your agent to list all active deployments.
More questions about Theta? See everything the Theta Connector can do
Connect Theta to Claude, Cursor, ChatGPT & more
