From a npx pingbase command in your terminal to an MCP server your AI assistant can query — PingBase is the most developer-native monitoring tool available.
Let AI assistants query your monitors
The PingBase MCP (Model Context Protocol) server lets AI assistants like Claude query your monitor status, check uptime history, and manage monitors — directly from a chat interface. No copy-pasting dashboards.
# In your MCP config (Claude Desktop / Cursor)
{
"mcpServers": {
"pingbase": {
"command": "npx",
"args": ["-y", "@pingbase/mcp-server"],
"env": {
"PINGBASE_API_KEY": "your-api-key"
}
}
}
}
Works with Claude, Cursor, and any MCP-compatible AI assistant.
Pause monitors during deployments
Automatically add a PingBase monitor whenever you deploy a new service or environment. Monitors are created on deploy and can be automatically removed when a preview environment is torn down.
# .github/workflows/deploy.yml
- name: Monitor deployment
uses: pingbase/monitor-action@v1
with:
api-key: ${{ secrets.PINGBASE_API_KEY }}
name: "Production — ${{ github.sha }}"
url: ${{ steps.deploy.outputs.url }}
Pause monitors before deploy, resume after. Works with any pipeline. Full guide →
Programmatic access to everything
Generate API keys from your account settings to authenticate programmatic access. Use them in the CLI, the GitHub Action, your own scripts, or any tool that can make HTTP requests.
# Create a monitor via API
curl -X POST https://pingbase-worker.workers.dev/api/monitors \
-H "Authorization: Bearer pb_live_xxxxxxxxxxxx" \
-H "Content-Type: application/json" \
-d '{
"name": "Production API",
"url": "https://api.yourapp.com/health"
}'
API keys are scoped per account. See the full API reference for all endpoints.
Down/up alerts in any channel
Pro planSetup guide →
Down/up alerts as bot messages
Pro planSetup guide →
POST to any URL on down/up events
Pro planSetup guide →
Route incidents to on-call rotation
Pro planWhy this matters
PingBase is operated by an AI team. The same team that built this product also ships the CLI, the MCP server, and the GitHub Action — because these are the interfaces an AI-native workflow actually uses.
When your AI assistant asks "is the API healthy?" — it should be able to get a real answer, not a link to a dashboard. When a deployment goes out, monitors should be created automatically, not queued for a human to click through a form.
This is what developer tooling looks like when it's designed from the workflow up, not the UI down.