Uptime monitoring MCP server

Why add checks from the editor

The best time to add a health check is when the endpoint is written. Usually it happens later, in a dashboard, after the first outage nobody noticed. With MCP you ask in the same session: “Watch the health URLs in this repo and put each one on our status page.” The agent finds the URLs, creates the checks, and links them to components.

Everything it creates is an ordinary check, page, or incident. The rest of the team sees the same thing in the console and the REST API.

What the agent can do

• List and create checks: Website, API, DNS, and TCP
• Pause, resume, or delete a check, and run a probe right now
• Read recent results for a check
• Create or update a public status page, its components, and component groups
• Open incidents and post updates
• List incident drafts from checks set to Ask me to review first, and publish or dismiss one when you ask
• Schedule, update, or cancel maintenance windows
• Start or disconnect Slack, and attach and test alert destinations you already created

The server has 34 tools. Cursor enables at most 40 MCP tools across all servers, so there is room left for your others.

What stays with people

Team invites and roles, creating and revoking API keys, billing, creating or deleting alert destinations, deleting an incident, and visitor subscribers stay in the console. There are no MCP tools for them, so the agent cannot give itself more access.

The server also tells the agent to check plan usage before creating anything, to stop at a cap, and to delete, open incidents, or email subscribers only when you asked in the conversation. On the free plan the API refuses a sixth check either way.

Set it up in two minutes

1. In the console, an owner or admin creates a key on API Keys and copies it once.
2. Export it in the shell your editor uses: export SP_API_KEY=sp_…
3. Install the skill and wire MCP:

npx skills add ScaledPapers/skills --skill uptime-status-page -a claude-code -a codex -g -y
bash ~/.agents/skills/uptime-status-page/scripts/wire-mcp.sh cursor   # or claude, or codex

The script merges your MCP config and stores a reference to SP_API_KEY, so the raw key never lands in a config file.

Or add it by hand

Cursor (~/.cursor/mcp.json):

{
  "mcpServers": {
    "scaledpapers": {
      "url": "https://api.scaledpapers.com/mcp",
      "headers": { "Authorization": "Bearer ${env:SP_API_KEY}" }
    }
  }
}

Claude Code (keep the single quotes, so it stores the variable and not the key):

claude mcp add --transport http scaledpapers https://api.scaledpapers.com/mcp \
  --header 'Authorization: Bearer ${SP_API_KEY}'

Codex (~/.codex/config.toml):

[mcp_servers.scaledpapers]
url = "https://api.scaledpapers.com/mcp"
bearer_token_env_var = "SP_API_KEY"

Checks that don't page you for one blip

Checks the agent creates behave like any other. One failed run is stored, not sent. The check is down only when failures in a row reach the confirmation period (1, 2, 3, or 5; default 3), and then one alert goes out. A check linked to a component can open an incident automatically, or put a draft up for review first. On Pro, checks run every 60 seconds from five regions: San Francisco, Ashburn, Montreal, Frankfurt, and Singapore.

Pricing

• Free: 5 checks, 5-minute interval, 1 region, 14 days of results, 1 status page
• Pro, $15/month: 25 checks, 60-second interval, 5 regions, 90 days of results, 5 status pages, private pages, custom domain, and Slack, Discord, PagerDuty, Pushover, and webhook alerts

API requests are unlimited and never billed. See pricing