Deploy a working Model Context Protocol (MCP) server in GCP.

MCP Server Blueprint

This blueprint includes a preconfigured MCP server, container (available in Python, Node or Golang), and gateway connection so you can deploy a secure, scalable MCP endpoint, hosted in GCP, in minutes.

Deploy This Blueprint

About the MCP Server Blueprint

This blueprint provides a working MCP server implementation. The server exposes tools through the Model Context Protocol and includes a sample weather tool for testing MCP workflows.

After deployment, MCP clients can connect to the server and invoke available tools through the exposed endpoint.

Use this blueprint as:

  • An MCP learning project
  • A starting point for custom MCP tools
  • A foundation for AI agents
  • A template for production MCP services
MCP Python blueprint graph
Gateway and container graph

How the Graph Works

This blueprint combines a gateway and a container that are already connected and configured, giving you a ready-to-deploy architecture that can be customized for your own tools and workflows. All components are deployed into a GCP solution with Shoal taking care of all the wiring and configuration.

Gateway

Receives incoming requests and exposes your MCP server through a secure public endpoint. It handles routing traffic to the container and provides the URL used by MCP-compatible clients.

Container

Runs the MCP server and hosts your MCP tools and business logic. Available in Python, Node.js, or Golang. Start with the included example or replace it with your own MCP server code.

What You Can Build with This Blueprint

Use the MCP Server blueprint to build AI-powered applications, connect existing systems, or create custom MCP tools that can be accessed by agents and MCP-compatible clients.

AI Agent Tools

AI Agent Tools

Give AI agents access to custom APIs, databases, or business logic.

Internal Company Tools

Internal Company Tools

Expose internal systems through MCP-compatible interfaces.

Data & Analytics Assistants

Data & Analytics Assistants

Allow agents to query reporting systems and data sources.

Automation Workflows

Automation Workflows

Create MCP tools that trigger backend processes or scheduled..

Support Assistants

Support Assistants

Connect AI agents to internal knowledge systems and workflows.

Dev & Engineering Tools

Dev & Engineering Tools

Connect AI agents to internal APIs and deployment workflows.

Using This MCP Blueprint

Move from code to a running backend with a visual setup flow that keeps services, connections, and deployment easier to manage.

1

Create a New Project

Create a new project in Shoal to get started.

2

Browse Blueprints

Open your environment and select Browse Blueprints to view available deployment templates.

3

Select the MCP Server Blueprint

Choose the MCP Server Blueprint to automatically generate the required components and connections.

4

Customize the MCP Server

Use the included weather example or replace it with your own MCP server code and tools.

5

Deploy

Click Deploy and let Shoal handle the cloud configuration, networking, and service wiring.

6

Connect Your Application

Use the generated MCP endpoint to start making MCP requests.

After Deployment

Once deployed, Shoal generates a live MCP endpoint that can be connected to Cursor, Claude Desktop, OpenAI Agents, and other MCP-compatible clients.

You can immediately test the included weather tool or upload your own MCP server implementation and redeploy.

MCP-compatible clients connected to the deployed endpoint

Related Blueprints

Move from code to a running backend with a visual setup flow that keeps services, connections, and deployment easier to manage.

Full-Stack Next.js ApplicationFull-Stack Next.js Application
Scheduler with Python ServiceScheduler with Python Service

Ready to Deploy Your MCP Server?

Start with a working MCP server blueprint, replace the example tools with your own code, and deploy a secure MCP endpoint without handling Google Cloud configuration manually.

Deploy This Blueprint