Service Virtualization MCP server

Virtual services emulate the behavior of unavailable or expensive web services to remove dependencies during development and testing. The MCP (Model Control Protocol) server for BlazeMeter service virtualization provides a local, on-premise solution for integrating AI-driven capabilities with the BlazeMeter service virtualization platform. Similar to existing MCP solutions for BlazeMeter, this server enables advanced automation and analysis features for creating and running virtual services.

This MCP provides:

  • AI-driven control of virtual services via MCP agents.

  • Context-aware simulation of dynamic, stateful APIs.

  • Natural language and low-code interfaces for managing mock services.

  • Intelligent orchestration of test flows using GenAI agents integrated with BlazeMeter service virtualization APIs.

This MCP makes service virtualization accessible to non-technical users through its conversational interface.

What you can do with the SV MCP Server

The service virtualization MCP server is built for developers and QA teams who want to enable AI agents, assistants, and chatbots to manage entire service virtualization workflows using natural language queries.

Through MCP integration with service virtualization, developers can:

  • Query a BlazeMeter service virtualization project, workspace, or account

  • Create, deploy, and stop virtual services

  • Create and assign HTTP transactions

  • Validate transactions in a sandbox

  • Create reusable virtual service templates

  • Manage keystores

Example prompt: 

Create a service "demo template service" in my active workspace.
Create a transaction with name "txn", method matcher POST, url /test
For the response, extract the value of the XML element name from the following request body:
    <product>
    <name>Book</name>
    </product>
The template must use Handlebars xPath helper and should correctly 
select the inner element value using an XPath expression.
The expected output from the template should be the value of the 
inner field (Book) when the request body matches the structure above.
Validate this template.
Also validate this transaction in the sandbox
Then create a virtual service VS1 in a US location, assign this transaction, and deploy it.

Prerequisites

Only Docker-based deployment is supported.

  • BlazeMeter API credentials: API Key ID and Secret. See BlazeMeter API keys.

  • Compatible MCP client (VS Code, Claude Desktop, Cursor, Windsurf, etc.)

  • Docker

Installation

For complete and current installation steps, see the GitHub README.

The SV MCP server supports two modes:

  • "type":"stdio"

    STDIO mode (default)

  • "type": "streamable-http"

    HTTP mode (streamable HTTP)

To learn more about where to find your API keys, see BlazeMeter API keys.

STDIO mode (default)

To configure the default STDIO mode, you need only the client config.

Client config

{
    "STDIO VS MCP": {
        "STDIO VS MCP": {
            "timeout": 60,
            "type": "stdio",
            "command": "docker",
            "args": [
                "run",
                "--rm",
                "-i",
                "-e",
                "API_KEY_ID=your_API_key_id",
                "-e",
                "API_KEY_SECRET=your_API_key_secret",
                "-e",
                "SOURCE_WORKING_DIRECTORY=/tmp",
                "-v",
                "/host/assets/path",
                "us-docker.pkg.dev/verdant-bulwark-278/sv-mcp/sv-mcp:latest"
            ]
        }
    }
}

HTTP mode (streamable HTTP)

To enable HTTP mode, set the environment variable MCP_MODE=http and expose port 8000.

Start the server first before the client connects to it.

Run

In the bash terminal , run:

docker run --rm -i -e MCP_MODE=http -e SOURCE_WORKING_DIRECTORY=/tmp -e API_KEY_ID=key -e API_KEY_SECRET=secret -p 8000:8000 --network host -v /host/assets/path:/tmp us-docker.pkg.dev/verdant-bulwark-278/sv-mcp/sv-mcp:latest

The server will be available at: http://localhost:8000/mcp

Client config

{
    "mcpServers": {
        "my-server": {
            "type": "streamable-http",
             "url": "http://localhost:8000/mcp"
         },
    }
 }

Enable or disable MCP tools

The MCP server supports enabling or disabling specific tools at startup using the environment variable MCP_ENABLED_TOOLS. If the environment variable is not set or is an empty string, all available tools are automatically enabled. To enable only certain tools, set the environment variable to a comma-separated list of tool names. For example: 

MCP_ENABLED_TOOLS="blazemeter_user,blazemeter_account,virtual_services_virtual_service"

For a list of tools, see Service Virtualization MCP Server tools reference.

Support and licensing