pingcheck-mcp

PingCheck MCP Server

An MCP server that lets any AI coding agent (Claude, Cursor, Copilot, Windsurf, Codex…) read your PingCheck status data.

Works with any MCP client — not just the ones named above. This server speaks the standard MCP protocol over stdio, so every AI coding tool that supports MCP can connect: Claude Desktop, Cursor, VS Code + GitHub Copilot, Cline, Roo, Windsurf, Zed, Trae, Codex, WorkBuddy, and more. There is no client allow-list — the only thing that changes is whether you supply a PINGCHECK_API_KEY (to read your own monitors) or omit it (public pages only).

• Public status pages: read with no API key. Status data is already public — just give the agent a slug.
• Your own monitors: authenticate with your pk_ API key (generate it in PingCheck Dashboard → API Key sidebar; admin accounts bypass the Pro/Team requirement) to list your pages, see component uptime, and trigger on-demand checks.

This server is an MCP server over the existing PingCheck REST API. Read tools work with or without an API key; write tools require your pk_ API key and only modify resources you own.

Three ways to provide your API key

You never have to edit a config file. Pick whichever is easiest:

#	Method	Best for	How
A	Tell the AI in chat ⭐	Non-technical users	After installing the server, just say "my PingCheck key is pk_...". The agent calls pingcheck_set_api_key, which verifies and stores it for the conversation. No file editing at all.
B	Env in config file	Persistent setup	Add PINGCHECK_API_KEY to the env block of your MCP config (see below). Survives restarts.
C	Let the AI edit the config	Agents with file access (WorkBuddy, Cursor…)	Say "add my PingCheck key pk_... to my MCP config" — the agent writes it to the file for you.

Method A is the recommended path if you're not comfortable finding and editing JSON config files. The key set via chat lives in memory for that session only; use B to make it permanent.

Tools (10 total)

Tool	Auth	What it does
pingcheck_public_status	none	Live status of any public page by slug
pingcheck_set_api_key	none	Set your pk_ key in chat — no config file needed
pingcheck_auth_status	none	Report whether you're authenticated ("am I connected?")
pingcheck_list_my_pages	pk_ key	List your own status pages
pingcheck_page_detail	pk_ key	Full detail (components + incidents) for one page
pingcheck_check_now	pk_ key	Force an on-demand re-check of a component
pingcheck_create_component	pk_ key	Add a new monitoring component to your page
pingcheck_delete_component	pk_ key	Delete a component from your page
pingcheck_create_incident	pk_ key	Report a new incident on your status page
pingcheck_update_incident	pk_ key	Update or resolve an existing incident

Install / run

# via npx (after publishing to npm)
npx pingcheck-mcp

# or clone + build
npm install
npm run build
node dist/index.js

Connect from your AI tool

✅ Easiest: install key-less, then set your key in chat (no JSON editing)

• Add PingCheck MCP to your client with no env block at all — the server runs fine without a key (public pages work immediately):

  { "mcpServers": { "pingcheck": { "command": "npx", "args": ["pingcheck-mcp"] } } }
  

• In your AI chat, say: "My PingCheck key is pk_xxx".
• The agent calls pingcheck_set_api_key, verifies it, and you now have full access — no config file touched. Ask "am I connected?" anytime to check.

One-click from the PingCheck dashboard (copies a ready config)

• Log in to pingcheck.cloud → Dashboard → API Key.
• Scroll to "Connect your AI agent".
• Pick your client (WorkBuddy / Claude Desktop / Cursor / VS Code / Zed…).
• Click Copy (Free: key-less config for public pages) or Generate & copy (Pro/Team: auto-creates your API key and copies a ready-to-paste config with it included).
• Paste into your client's config file. Done.

Manual config (advanced)

PingCheck MCP uses the standard mcpServers config block. Paste the snippet below into whatever MCP client you use — the JSON is identical across Claude Desktop, Cursor, VS Code (Copilot), Cline, Roo, Windsurf, Zed, Trae, Codex, WorkBuddy, and any other MCP-compatible tool. Only the location of the config file differs from client to client.

{
  "mcpServers": {
    "pingcheck": {
      "command": "npx",
      "args": ["pingcheck-mcp"],
      "env": {
        "PINGCHECK_API_KEY": "pk_xxx_your_key_here"
      }
    }
  }
}

Where to paste it (common clients):

Client	Config location
Claude Desktop	claude_desktop_config.json (OS config dir)
Cursor	Settings → MCP, or ~/.cursor/mcp.json
VS Code (Copilot)	.vscode/mcp.json in your workspace
Cline / Roo	Cline's MCP settings (cline_mcp_settings.json)
Windsurf	~/.codeium/windsurf/mcp.json
Zed	settings.json → context_servers
Trae	MCP settings panel
WorkBuddy	~/.workbuddy/mcp.json

Don't see your tool? If it supports MCP and lets you add a mcpServers entry, it works. The server itself imposes no client restrictions — it just speaks standard MCP over stdio.

Leave PINGCHECK_API_KEY out entirely if you only need public pages.

Example prompts once connected

• "My PingCheck key is pk_xxx" — authenticate in one line, no config file
• "Am I connected to PingCheck?" — check auth status
• "Is the Acme API up? Check pingcheck.cloud/status/acme"
• "List my PingCheck pages"
• "Re-check my database component now"

Notes

• Defaults to https://pingcheck.cloud; override with PINGCHECK_BASE_URL.
• Public reads need no token; private reads require a pk_ key (Pro/Team).
• Rate limits apply on the PingCheck side; the agent should not hammer.

Publish to the MCP Registry

npm install -g mcp-publisher
mcp-publisher login github     # device OAuth, personal account, no company
mcp-publisher publish          # zero-review publish to the official Registry

Then list it on community directories (Glama, mcp.so) for discoverability. The mcpName field in package.json must match io.github.Churman1113/pingcheck-mcp.