If you are anything like me, you spend most of your day in a terminal. Switching to a browser dashboard to check whether your API is still responding feels like an interruption — a small one, but it adds up. What if you could ask your AI coding assistant to check on your monitors, create new ones, or post an incident update, all without leaving your editor?

That is exactly what CronAlert's MCP server enables. Using the Model Context Protocol, Claude Code can talk directly to your CronAlert account — listing monitors, pulling uptime stats, managing incidents, and more. This post walks through how to set it up and what you can actually do with it.

A quick primer on Claude Code

Claude Code is Anthropic's CLI tool that brings Claude into your terminal. You install it globally with npm: 

npm install -g @anthropic-ai/claude-code

Then run claude in any directory to start a session. It can read your files, run commands, write code, and — crucially for this article — connect to external services through MCP servers. If you have used GitHub Copilot in the terminal or Cursor's AI features, Claude Code is in the same category, but with the added ability to plug into arbitrary APIs via MCP.

What is the Model Context Protocol (MCP)?

MCP is an open standard that defines how AI tools connect to external services. Think of it like a USB-C port for AI: a single protocol that lets any compliant AI client talk to any compliant server. Instead of copy-pasting curl output into a chat window, the AI calls tools directly and gets structured data back.

CronAlert publishes an MCP server as an npm package (cronalert-mcp). Once registered with Claude Code, it gives Claude direct access to 11 monitoring tools — everything from creating monitors to posting incident updates.

Prerequisites

You need four things:

  • Node.js 18+ — the MCP server runs via npx
  • Claude Code — installed globally (npm install -g @anthropic-ai/claude-code)
  • A CronAlert account — the free plan works fine, no credit card needed
  • An API key — generate one at Settings > API Keys

Setup: one command

The entire setup is a single line. Open your terminal and run: 

claude mcp add cronalert -e CRONALERT_API_KEY=YOUR_API_KEY -- npx -y cronalert-mcp

That is it. Here is what happens behind the scenes:

  • claude mcp add tells Claude Code to register a new MCP server
  • cronalert is the name you are giving this server (you will see it in /mcp listings)
  • -e CRONALERT_API_KEY=... passes your API key as an environment variable to the server process
  • npx -y cronalert-mcp is the command Claude Code will run to start the server — npx downloads the package on first use, and -y skips the confirmation prompt

The server runs locally on your machine via stdio transport. Your API key never leaves your computer except when making authenticated requests to the CronAlert API.

Verify the connection: Start a Claude Code session and type /mcp. You should see cronalert listed with a green status. If it shows as disconnected, double-check that your API key is correct and that npx cronalert-mcp runs without errors on its own.

What you can actually do with it

The real value is not in the setup — it is in how this fits into your workflow. Here are concrete examples of prompts you can give Claude Code and what happens under the hood.

Create a monitor while you are deploying

You just finished setting up a new service. Instead of opening the CronAlert dashboard, you type: 

# In your Claude Code session:
Create a monitor for https://api.myapp.com/health that checks every 3 minutes

Claude calls the create_monitor tool with the URL and interval. The monitor starts checking immediately. You stay in your terminal, move on to the next task.

Check what is down right now

List all my monitors and highlight any that are currently down

Claude calls list_monitors, gets back the full list with statuses, and presents a summary. No dashboard, no page load, no clicking through filters.

Pull uptime stats for a specific service

What's the uptime percentage and average response time for my production API this week?

This hits get_check_results with the appropriate monitor ID and time range. Claude parses the response and gives you the numbers in plain English — uptime percentage, average response time, any incidents that occurred.

Batch operations during maintenance

Pause all monitors that have "staging" in the name

Claude calls list_monitors first to find the relevant ones, then calls update_monitor on each to set them to paused. When your maintenance window is over, tell it to resume them. This kind of batch operation is tedious in any dashboard but trivial with natural language.

Post an incident update

Post an update to the active incident on api.myapp.com saying we identified the
root cause as a database connection pool exhaustion and a fix is being deployed

Claude finds the active incident via list_incidents, then calls add_incident_update with your message. Your status page subscribers get notified. You wrote it from the same terminal where you are fixing the bug.

Why this beats switching to a dashboard

There are a few reasons this workflow sticks once you try it:

  • Zero context switching. You are already in the terminal writing code. Checking uptime or creating a monitor is just another prompt in the same session.
  • Deployment + monitoring in one flow. Deploy your service, create a monitor for it, and verify it is responding — all in a single Claude Code session without touching a browser.
  • Natural language over forms. "Create monitors for these 5 URLs with 1-minute intervals" is faster than filling out a form five times. Claude handles the repetition.
  • Composability. Combine CronAlert's MCP with other MCP servers. If you have GitHub's MCP server connected, you could merge a PR, deploy, and set up monitoring in one conversation.

Tips for power users

Verify your connection with /mcp

If something feels off — Claude is not recognizing monitoring commands, or you are getting authentication errors — run /mcp inside your Claude Code session. It shows the status of all connected MCP servers and lets you restart them if needed.

Plan limits still apply

The MCP server uses the same API as the dashboard, so your plan limits are enforced server-side. On the free plan, you get 25 monitors with 3-minute check intervals. If you need 1-minute intervals or more monitors, upgrade to Pro ($5/month). The MCP server itself is free to use on any plan.

Combine with other MCP servers

This is where things get interesting. If you have GitHub, Sentry, or other MCP servers registered in Claude Code, you can build cross-service workflows in a single conversation. For example: "Check if there are any open incidents on my API, look at the recent Sentry errors for that service, and draft a post-mortem." Each MCP server contributes its piece.

Available tools reference

The CronAlert MCP server exposes 11 tools. Here is the full list:

Tool Description
list_monitors List all monitors with filtering and pagination
create_monitor Create a new HTTP monitor with URL, interval, and alert settings
get_monitor Get detailed info for a specific monitor including current status
update_monitor Update monitor configuration, or pause/resume checking
delete_monitor Permanently delete a monitor and its check history
get_check_results Retrieve check history with uptime percentage and response times
get_monitor_incidents Get all incidents for a specific monitor
list_incidents List all active incidents across your account
list_status_pages List your public status pages
add_incident_update Post a status update on an open incident
get_incident_updates Get the full timeline of updates for an incident

Every tool accepts and returns JSON. Claude handles the serialization — you just describe what you want in plain English.

Frequently asked questions

Do I need a paid CronAlert plan to use the MCP server?

No. The free plan gives you 25 monitors with 3-minute check intervals, email/Slack/Discord/webhook alerts, a status page, and full API access including MCP. Paid plans unlock 1-minute intervals, more monitors, and features like keyword monitoring and maintenance windows.

Does the MCP server work with Claude Code on all platforms?

Yes. Claude Code runs on macOS, Linux, and Windows. The MCP server is a Node.js process started via npx, so it works anywhere Node.js 18+ is installed.

Can I use the remote endpoint instead of the npm package?

CronAlert also offers a remote MCP endpoint at https://cronalert.com/mcp using Streamable HTTP transport. It works with any MCP-compatible client that supports remote servers. However, for Claude Code specifically, the npm package (cronalert-mcp) is recommended because it runs locally via stdio, which avoids network latency on every tool call.

What happens if the MCP server disconnects?

Claude Code manages MCP server lifecycle automatically and will attempt to reconnect. If you suspect a connection issue, type /mcp in your session to check the status of all registered servers. You can restart a specific server from that menu.

Wrapping up

Setting up CronAlert's MCP server in Claude Code takes about 30 seconds and removes the need to context-switch to a dashboard for routine monitoring tasks. Creating monitors, checking uptime, pausing checks during maintenance, posting incident updates — all of it becomes a natural language prompt in your terminal.

If you do not have a CronAlert account yet, sign up for free and generate an API key. Then run the one-liner: 

claude mcp add cronalert -e CRONALERT_API_KEY=YOUR_API_KEY -- npx -y cronalert-mcp

Your monitors are now a conversation away.