Using AI with Oh Dear

AI coding assistants like Claude Code can interact with Oh Dear on your behalf. Instead of switching to the dashboard to check uptime, investigate broken links, or start a maintenance window, you can simply ask your AI assistant to do it for you.

For example, you could ask:

• "Are any of my monitors down right now?"
• "Show me the broken links on example.com"
• "Start a maintenance window for 30 minutes on my API server"
• "What are the Lighthouse scores for my homepage?"

Oh Dear provides two ways to make this work: the CLI skill and the MCP server.

CLI skill (recommended) #

The CLI skill is the easiest way to get started. It teaches your AI assistant how to use the Oh Dear CLI tool, giving it access to every Oh Dear API endpoint. Install it with a single command:

ohdear install-skill

This is the recommended approach because it covers the full Oh Dear API, requires no additional configuration beyond the CLI itself, and works with any AI assistant that supports the Skills protocol.

MCP server #

The MCP server connects AI assistants directly to your Oh Dear data through the Model Context Protocol. It provides a curated set of read-focused tools for querying monitors, uptime, broken links, DNS history, Lighthouse reports, and more.

This is a good option when you cannot install the CLI, or when your AI client supports MCP but not skills.