MCP servers for DevOps: drive your infra from Cursor and Claude Code

Your IDE Already Knows How to Code. Now Teach It Your Infrastructure.

MCP (Model Context Protocol) is the open standard that lets AI coding tools — like Cursor, Claude Code, and Codex — talk to external systems. Instead of copying logs into a chat window, your AI assistant can query your infrastructure directly.

IAN ships as an MCP server. That means you can audit repos, check cloud costs, and trigger deployments without leaving your editor.

What MCP Actually Is

MCP is a protocol (think: USB for AI tools) that defines how AI assistants discover and use external tools. An MCP server exposes capabilities — “tools” — that the AI can call when relevant.

When you connect IAN as an MCP server, your AI coding environment gains these capabilities:

• audit_repository — run a security and cost audit on any connected repo
• get_findings — retrieve current audit findings with severity and remediation suggestions
• cost_summary — get cloud cost breakdown by service, account, or team
• deploy_status — check the status of recent deployments
• create_remediation — generate a fix PR for a specific finding

See the IAN team run on your cloud. We connect to your AWS account via a scoped read-only role, run the Observe-tier agents, and leave you with a concrete audit report — cost waste, security exposure, compliance gaps, and a labor-offset estimate. You keep the findings regardless of next steps. Get a free infrastructure audit →

Setting It Up in Cursor

Add IAN to your Cursor MCP configuration:

{
  "mcpServers": {
    "ian": {
      "command": "ian",
      "args": ["mcp", "serve"],
      "env": {
        "IAN_API_TOKEN": "your-api-token"
      }
    }
  }
}

Once connected, you can ask Cursor things like:

• “What are the critical findings in the payments-api repo?”
• “How much are we spending on RDS this month?”
• “Generate a fix for finding #247”
• “What’s the deployment status for staging?”

Cursor calls the IAN MCP server, gets the data, and responds in context.

Setting It Up in Claude Code

Claude Code supports MCP servers natively. Add IAN to your project’s .mcp.json:

{
  "servers": {
    "ian": {
      "command": "ian",
      "args": ["mcp", "serve"]
    }
  }
}

Now Claude Code can pull infrastructure context directly into your coding session. Ask it to fix a security finding, and it can read the finding details, understand the codebase, and generate the fix — all in one flow.

Why This Matters

The traditional workflow for handling a security finding looks like:

1. Open security dashboard
2. Read the finding
3. Switch to IDE
4. Find the relevant code
5. Understand the vulnerability
6. Write the fix
7. Test it
8. Create a PR

With MCP, it becomes:

1. Ask your AI assistant: “Fix the critical findings in this repo”
2. Review the generated changes
3. Commit

The context switching disappears. The AI has access to both the finding details (via IAN) and the codebase (via your editor). It connects the dots.

Beyond Audits: Infrastructure as Context

MCP isn’t just for security. When your AI assistant has access to your infrastructure context, it makes better decisions about everything:

• Writing Terraform? It can check your current resource usage first.
• Debugging a performance issue? It can pull recent cost anomalies and deployment logs.
• Reviewing a PR? It can check if the change introduces compliance drift.

Your infrastructure becomes part of the AI’s context window — always available, never stale.

Try It

Install the IAN CLI, connect your repos, and add IAN as an MCP server in your editor. Your AI coding assistant just became your AI DevOps assistant too.

Get a free infrastructure audit → | See pricing →