Skip to main content
The Alertee MCP (Model Context Protocol) server lets AI assistants — like Claude, Codex, and Cursor — create and manage Alertee checks as part of their workflow.

The use case

You’re debugging a bug in another project’s session. You notice a pattern that could cause problems in production. Instead of switching context to open the Alertee dashboard, you just ask your assistant:
“Create an Alertee check that alerts if failed payments exceed 10 in the last hour.”
It calls the create_check tool, connects to your organization, and creates the check — all without you leaving your current session. It also works the other way: an alert can fire a webhook that hands an AI agent the check, and the agent uses MCP to pull the SQL, recent history, and failing data and investigate on its own. See Investigate alerts with an AI agent.

Two ways to connect

Most people should use the remote server — there’s nothing to install. If you already live in the CLI, the local server reuses your existing login. Both expose the same tools. Connect a client →

How the remote server works

You add the URL to your AI client and it opens a browser window to sign in to Alertee and approve access — the same OAuth flow you’d expect from any “Connect your account” button. Once approved, the client connects and can call tools on your behalf, against your default organization. Nothing to install.

How the local server works

The local server runs alongside your client and reuses your Alertee CLI login — no extra setup once you’ve run alertee login and alertee link.

Available tools

See the tools reference for full parameter details.

Next step

Connect a client →