Pulumi MCP Server

What is this?

The Pulumi MCP Server is a specialized implementation of the Multi-Channel Protocol (MCP) that provides AI-driven agents and developer workflows with a robust, programmable interface to Pulumi environments. Pulumi is a leading Infrastructure as Code platform enabling teams to define, deploy, and manage cloud resources using familiar programming languages. By exposing Pulumi’s capabilities via an MCP server, it bridges the gap between conversational AI agents and real-world infrastructure operations.

Within the broader MCP ecosystem, the Pulumi MCP Server focuses on infrastructure automation, allowing AI assistants to discover, preview, deploy, and inspect Pulumi stacks programmatically. Organizations building advanced DevOps pipelines, AI-driven CI/CD bots, or autonomous remediation systems can leverage this server to reduce manual toil, accelerate deployment cycles, and enforce secure, declarative control over resources.

Quick Start

Install the server using npm:

npm install @modelcontextprotocol/server-pulumi-mcp-server

Then add it to your MCP client configuration:

{
  "mcpServers": {
    "pulumi-mcp-server": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-pulumi-mcp-server"],
      "env": {
        "API_KEY": "your-api-key-here"
      }
    }
  }
}

Key Features

Feature 1: Automated IaC Operations: AI agents can programmatically initiate Pulumi previews, updates, and state inspections to reduce manual toil and accelerate deployment cycles.

Feature 2: Secure State Management: Leverages Pulumi Cloud API authentication and fine-grained access controls for secure stack operations.

Feature 3: Flexible Integration: Interoperates seamlessly with chatbots, CI/CD pipelines, and custom AI assistants built on the MCP framework.

Example Usage

For example, you can list all stacks in your project with a single call:

// Example code
const result = await client.callTool({
  name: "listStacks",
  arguments: {
    organization: "acme-corp",
    project: "web-app"
  }
});

This code invokes the listStacks tool and returns the available stacks in your Pulumi project.

Configuration

The server accepts the following environment variables:

API_KEY – Your Pulumi access token for authenticating with the MCP server.

SETTING_1 (optional) – The port the server listens on (default: 8080).

Available Tools/Resources

listStacks: Retrieve names and metadata of all stacks in an organization or project.

previewUpdate: Run a dry-run of resource changes and return a detailed change plan.

performUpdate: Execute deployments with optional auto-approval.

Who Should Use This?

This server is perfect for:

Use case 1: DevOps teams and CI/CD engineers looking to automate infrastructure workflows and integrate with AI assistants.

Use case 2: Software developers who want programmatic control over Pulumi stacks from chatbots or scripts.

Use case 3: AI practitioners building autonomous agents for infrastructure provisioning and management.

Conclusion

The Pulumi MCP Server unlocks AI-driven infrastructure automation by providing a standardized MCP interface for stack lifecycle operations, previews, updates, and outputs. Install the npm package, configure your credentials, and register the server to start building intelligent DevOps workflows today.

Check out the GitHub repository for more information and to contribute.