How to Monitor Website Uptime from Claude Desktop with MCP

Not everyone lives in the terminal. If you spend your day in a graphical desktop environment and would rather type a question than navigate a dashboard, Claude Desktop with CronAlert's MCP server is a surprisingly effective way to manage your uptime monitoring. You get a conversational interface where you can create monitors, investigate incidents, check uptime stats, and pause checks during maintenance windows — all by asking Claude in plain English.

This guide walks through the full setup: installing the MCP server, connecting it to Claude Desktop, and the kinds of tasks you can hand off to it once everything is running.

A quick primer on Claude Desktop and MCP

Claude Desktop is Anthropic's native application for macOS and Windows. It gives you a rich chat interface with Claude — file uploads, image analysis, conversation history, and critically for our purposes, support for MCP servers. If you've used Claude through the web, this is the same model with a native app wrapper and some extra capabilities.

MCP (Model Context Protocol) is an open standard that lets Claude connect to external tools and services. Think of it as giving Claude structured permission to call an API on your behalf. When you add CronAlert as an MCP server, Claude gains the ability to list your monitors, create new ones, pull check history, post incident updates, and more — all through the same chat window where you'd ask any other question.

Prerequisites

Before you start, make sure you have:

• Claude Desktop installed (claude.ai/download) with a Pro, Max, or Team subscription
• Node.js 18 or later installed on your system (the MCP server runs via npx)
• A CronAlert account — the free plan works fine for this
• A CronAlert API key, which you'll generate in the next step

Setup in five steps

Head to cronalert.com/app/settings/api-keys and click Create API Key. Give it a descriptive name like "Claude Desktop" so you can identify it later. Copy the key — you'll only see it once.

Claude Desktop stores its MCP server configuration in a JSON file on your local machine. The location depends on your operating system:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

You can also reach this file from inside Claude Desktop: open Settings, go to the Developer tab, and click Edit Config. That will open the file in your default text editor.

Paste the following configuration into your claude_desktop_config.json. If the file already has other servers, add the "cronalert" key inside the existing mcpServers block:

{
  "mcpServers": {
    "cronalert": {
      "command": "npx",
      "args": ["-y", "cronalert-mcp"],
      "env": {
        "CRONALERT_API_KEY": YOUR_API_KEY
      }
    }
  }
}

Replace YOUR_API_KEY with the key you generated in step 1. The -y flag tells npx to automatically install the cronalert-mcp package without prompting.

Close and reopen Claude Desktop. MCP server configuration is only read at startup, so a restart is required for changes to take effect.

In a new Claude Desktop conversation, look at the chat input area. You should see a hammer icon (the MCP tools indicator). Click it, and you'll see cronalert listed with its available tools. If it's there, you're ready to go.

Try a quick test: type "List my CronAlert monitors" and send it. Claude will call the list_monitors tool and show you what's configured on your account.

What you can do with it

The real value of this setup is that you interact with your monitoring infrastructure through conversation instead of clicking through a UI. Here are some practical examples — these are actual prompts you can type into Claude Desktop.

Check on your monitors

You can ask broad questions and let Claude figure out which tools to call:

"Are any of my monitors currently down?"

"Show me the uptime stats for my main website over the last week."

Claude will pull the relevant data, summarize it, and flag anything that needs attention. No navigating between dashboard tabs.

Create and manage monitors

"Create a new monitor for shop.mycompany.com — check it every minute."

"We're doing maintenance on the staging server tonight. Pause that monitor until I say otherwise."

Creating a monitor through the chat takes about ten seconds of typing versus filling in a form. Claude will confirm the details before creating it, so you can catch mistakes.

Investigate incidents

"There's an active incident on the API monitor. Post an update saying we're investigating a database connection issue."

"Show me all the incidents from the past month for the checkout service."

This is where the conversational interface really shines during an outage. You can post status updates, check incident timelines, and coordinate your response — all without leaving a single window.

The conversational advantage

A dashboard shows you data. A conversation lets you ask follow-up questions about it. That distinction matters more than it sounds.

Say you notice your API endpoint monitor has had three failures this week. On a dashboard, you'd click into the monitor, scroll through check results, and try to spot a pattern yourself. With Claude Desktop, you can just ask:

"Why did my API monitor fail 3 times this week? Is there a pattern?"

Claude will pull the check results, examine the timestamps and response codes, and give you an actual analysis. Maybe all three failures happened between 2 and 3 AM UTC during your deployment window. Maybe the response times were climbing for an hour before each failure. That kind of synthesis is something a traditional monitoring dashboard leaves you to do manually.

Troubleshooting

MCP tools not showing up in the chat input

First, make sure Node.js 18+ is installed by running node --version in a terminal. If the version is below 18, update it. Next, double-check that your config JSON is valid — a missing comma or bracket will silently prevent the server from loading. You can paste the file contents into any JSON validator to check. Finally, fully quit and restart Claude Desktop (not just close the window).

Authentication errors when calling tools

This usually means the API key is incorrect or has been revoked. Go to your API key settings and verify that the key is still active. If in doubt, generate a new one, update the config file, and restart Claude Desktop.

Windows-specific path issues

On Windows, the config path uses %APPDATA%, which typically resolves to C:\Users\YourName\AppData\Roaming. If you're editing the file manually, make sure you're in the right directory. The file is Claude\claude_desktop_config.json inside that folder. Also verify that npx is available in your system PATH — you can test with npx --version in PowerShell.

Available tools reference

CronAlert's MCP server exposes 11 tools. Here's the full list:

You don't need to memorize these. Claude knows what tools are available and will pick the right one based on what you ask. If you say "pause my staging monitor," Claude will use update_monitor without you needing to specify the tool name.

Frequently asked questions

Do I need a paid Claude plan to use MCP servers?

Yes. MCP server support in Claude Desktop requires a Claude Pro, Max, or Team subscription. On the CronAlert side, any plan works — including the free tier with 25 monitors.

Is my API key stored securely?

Your API key lives in a local configuration file on your machine. It's only read by the Claude Desktop application and is not uploaded to Anthropic's servers. That said, treat the config file like you would any file containing credentials — keep your machine locked and don't share the file.

Can I use this alongside the CronAlert dashboard?

Absolutely. The MCP integration and the web dashboard use the same API and operate on the same data. Create a monitor in Claude Desktop, and it shows up in the dashboard immediately. There's no conflict between the two — use whichever is more convenient for the task at hand.

What's the difference between this and the Claude Code setup?

Claude Desktop is a graphical chat application. Claude Code is a command-line tool for developers who prefer working in the terminal. Both use the same cronalert-mcp package and expose identical monitoring tools. The only difference is the interface. If you work in a terminal all day, Claude Code might feel more natural. If you prefer a visual app with conversation history you can scroll back through, Claude Desktop is the better fit.