- READY
- ACTIVE
- RUNNING
- HUMAN
Short answer
How can I run inference tasks using my AI?
Tell your agent to find a suitable system image and a specific GPU cluster, then have it launch the task. You'll get a breakdown of available resources and the ability to trigger execution immediately.
Deployment outcomes
Where your tasks land.
The agent maps your request to the right hardware and software.
READY
Resource visibility
Your agent identifies which system images are ready for use. You get a clear view of what software environments are available.
ACTIVE
Cluster selection
The agent maps out available GPU clusters. You can pick exactly where your workload needs to run based on hardware availability.
RUNNING
Task execution
The agent triggers the actual computing task. Your workload moves from a request to a live process on the hardware.
HUMAN
Manual oversight
You decide when to pull the trigger on a specific cluster. The AI handles the discovery, but you maintain control over the final deployment.
The workflow
What your AI does when the task arrives.
The agent handles the heavy lifting of resource discovery and command execution.
Scan images
The agent looks through the registry to see which system images are currently available for your workload.
list_imagesLocate hardware
It checks the available GPU clusters to find the right hardware match for your requirements.
list_clustersLaunch workload
Once the image and cluster are identified, the agent sends the command to start the computing task.
create_taskConfirm deployment
The agent verifies that the task was successfully sent to the cluster and reports the status back to you.
create_task
Try it
Copy these to start.
Use these prompts to begin your deployment.
Starting points
These are basic starting lines. Swap out the image names or cluster IDs with your specific requirements.
Shunwang Tech Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_bgiyuAWhRzWrIFnn1swCL5IcEpCA8FYUUwp4FNb7/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 — Shunwang Tech capabilities are ready to use.
{
"mcpServers": {
"shunwang-tech-mcp": {
"url": "https://edge.vinkius.com/vk_preview_bgiyuAWhRzWrIFnn1swCL5IcEpCA8FYUUwp4FNb7/mcp"
}
}
}Show me all the available system images.
Which GPU clusters are currently online?
Deploy a new task using the 'resnet-v2' image on the 'gpu-cluster-alpha' cluster.
Find an available cluster and launch a task with the 'llama-3-inference' image.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Shunwang Tech once, then ask.
One connection gives your agent access to your GPU resources. Everything stays secure and ready for your next command.
Connect Shunwang Tech to your AIFAQ
How this task behaves.
- 01
Can the AI delete existing system images?
No. The agent can only list available images; it cannot delete or modify them.
- 02
Does the AI choose the cluster automatically?
It can, but you should specify a cluster if you have strict hardware requirements. The agent follows your direction.
- 03
Can I monitor the task progress through the AI?
The agent can trigger the task, but real-time telemetry depends on your specific cluster configuration.
- 04
Will the AI charge my account for every list command?
No. Listing images and clusters is a read-only action that doesn't trigger task costs.
- 05
Can the AI change the GPU hardware specs?
No. The agent can only select from the clusters that already exist in your environment.
More questions about Shunwang Tech? The Connector page answers them. See everything the Shunwang Tech Connector can do
Connect Shunwang Tech to Claude, Cursor, ChatGPT & more
