How to Manage Cloud Infrastructure with Open WebUI Using the Hyperstack MCP Server on AI Studio

AI assistants like Claude Desktop and modern agent frameworks are changing how developers interact with AI. Instead of manually writing API calls, you can simply describe what you want in plain English and let AI carry out the task.

And how does this happen? The Hyperstack MCP (Model Context Protocol) Server makes it possible.

By connecting an AI client such as Open WebUI or Claude Desktop to Hyperstack through MCP, you can manage cloud infrastructure using natural language. The MCP Server translates AI requests into secure Hyperstack API operations. That means you can create resources, scale infrastructure and run operational workflows simply by asking what you need.

In this tutorial, you’ll learn how to install the Hyperstack MCP Server, connect it to Open WebUI and start managing your infrastructure using AI prompts.

Important Note: Live Infrastructure Changes

The MCP Server interacts with real Hyperstack resources. This means it can:

• Create infrastructure
• Modify resources
• Delete existing workloads

These operations may incur charges on your account. You must review actions carefully before confirming them.

*NexGen Cloud is not responsible for unintended infrastructure usage, billing charges or data loss resulting from the use of this tool.

What Will You Learn

By the end of this tutorial, you will be able to:

• Run the Hyperstack MCP Server locally
• Connect an AI client to Hyperstack infrastructure
• Execute infrastructure operations using natural language
• Use AI to automate multi-step infrastructure workflows

What is the Hyperstack MCP Server?

The Hyperstack MCP Server connects AI clients to the Hyperstack cloud platform using the Model Context Protocol (MCP). It allows AI assistants to interact with infrastructure in a structured and secure way.

Instead of manually writing API calls, MCP enables AI systems to use predefined tools that perform authenticated infrastructure operations on Hyperstack.

With the MCP Server running, AI clients can:

• Create and manage Virtual Machines
• Provision Kubernetes clusters
• Create and attach Volumes
• Retrieve billing and usage information
• Manage environments
• Execute multi-step infrastructure workflows

Prerequisites

Before starting, ensure the following:

• Docker (Installed and running)

• Port 8080 is available

• You have a valid Hyperstack API Key from the Hyperstack Console

Now we will walk you through the tutorial in two parts:

1. Running the MCP Server
2. Connecting via Open WebUI 

Part 1: Run the Hyperstack MCP Server

Follow the steps below to run the server:

Step 1: Run the server locally using Docker. 

docker run --rm \
  --name hyperstack-mcp \
  -p 8080:8080 \
  -e HYPERSTACK_API_KEY=your_api_key_here \
  ghcr.io/nexgencloud/hyperstack-mcp-server:latest

Important: The environment variable HYPERSTACK_API_KEY is required and must contain a valid Hyperstack API key.

Step 2: Verify the Server is Running

Once the container starts, verify that the MCP server is healthy.  Run:

curl http://localhost:8080/health  

The expected response would be:

{
  "status": "ok",
  "service": "hyperstack-mcp-server",
  "version": "0.1.0"
}

If the server does not respond, ensure Docker is running and the container started correctly.

Part 2: Connect via Open WebUI

Open WebUI provides a simple browser interface for interacting with AI models and MCP tools.

Step 1: Start Open WebUI

docker run -d \
  --network host \
  -e PORT=3000 \
  --name open-webui \
  ghcr.io/open-webui/open-webui:main

Once started, Open WebUI will be accessible at http://0.0.0.0:3000.

Step 2: Connect the MCP Server to Open WebUI

Next, configure Open WebUI to connect to the MCP Server.

1. Open Admin Panel

2. Navigate to Settings → External Tools

3. Click + to add a new connection

4. Change the Type to MCP Streamable HTTP

5. Set the URL to: http://127.0.0.1:8080/mcp  

6. Set Authentication to None

7. Assign a Name and ID

8. Click Save

Open WebUI can now communicate with the MCP Server. 

Step 3: Add a Hyperstack AI Studio Model

Next, configure Open WebUI to use a model from Hyperstack AI Studio.

1. Navigate to Admin Panel → Settings → Connections

2. Under OpenAI API, click +

3. Set the Server URL to https://console.hyperstack.cloud/ai/api/v1

4. Add your Hyperstack API key in the header: Authorization: Bearer <YOUR_API_KEY>
   

(Optionally, you can add specific Model IDs.)

Once the above steps are complete, click Save. Now the model will appear in the Chat interface.

Step 4: Enable MCP Tools in a Chat Session

Now enable the MCP tools for the session.

1. Open New Chat
2. Toggle Controls
3. Set Function Calling to Native
4. Open Integrations → Tools
5. Enable the MCP Server connection

Your AI client can now call Hyperstack infrastructure tools.

Step 5: Test the Integration

You can now manage infrastructure using natural language. Try prompts such as:

• "List all virtual machines in my account."
• "Show current billing information."
• "What environments are available?"

If the integration is configured correctly, Open WebUI will call the MCP tools through the Hyperstack MCP Server and return the results directly in chat.

Try MCP with Hyperstack AI Studio

Ready to see AI-driven infrastructure management in action?

With Hyperstack AI Studio, you can run powerful AI models that connect directly to the Hyperstack MCP Server and start managing cloud resources using natural language.

FAQs

What is the Hyperstack MCP Server?

The Hyperstack MCP Server is a service that connects AI clients to Hyperstack infrastructure through the Model Context Protocol (MCP). It translates structured AI tool calls into authenticated Hyperstack API operations, allowing infrastructure to be managed using natural language.

What is MCP (Model Context Protocol)?

The Model Context Protocol is an open protocol that enables AI models to interact with external tools and services in a structured way. MCP allows AI systems to safely execute operations such as querying databases, running workflows or managing infrastructure.

What infrastructure tasks can AI perform through the MCP Server?

Once connected, AI clients can perform tasks such as creating and managing virtual machines, provisioning Kubernetes clusters, attaching storage volumes, retrieving billing and usage data, managing environments, and executing multi-step infrastructure workflows.

Does the MCP Server modify real infrastructure?

Yes. The MCP Server interacts with real Hyperstack resources. This means it can create, modify, or delete infrastructure. Always review operations carefully to avoid unintended changes or unexpected charges.

Can MCP be used for AI agents and automation workflows?

Yes. MCP is designed to enable AI agents to execute structured tasks reliably. By exposing infrastructure operations as MCP tools, developers can build automated workflows where AI agents provision resources, monitor systems, and orchestrate deployments.

Where can I run AI models for this setup?

You can run models through Hyperstack AI Studio, which provides access to production-ready AI models that integrate with MCP-enabled tools and infrastructure services.