MCP Integration
Connect AI assistants like Claude and Cursor to CronSignal using MCP (Model Context Protocol). Create and manage monitors, debug failed jobs, and get intelligent suggestions - all through natural conversation.
Quick Start
1. Get Your API Key
Go to cronsignal.io/mcp to create an API key. You'll need this to authenticate the MCP server.
2. Configure Your AI Assistant
Cursor
Add to your Cursor settings (Settings → Features → MCP):
{
"mcpServers": {
"cronsignal": {
"command": "npx",
"args": ["cronsignal-mcp"],
"env": {
"CRONSIGNAL_API_KEY": "cs_live_your_key_here"
}
}
}
}
Claude Desktop
Add to your Claude config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"cronsignal": {
"command": "npx",
"args": ["cronsignal-mcp"],
"env": {
"CRONSIGNAL_API_KEY": "cs_live_your_key_here"
}
}
}
}
3. Restart Your AI Assistant
Close and reopen Cursor or Claude Desktop to load the MCP server.
Available Tools
Monitor Management
| Tool | Description |
|---|---|
list_checks |
List all monitors with their current status |
get_check |
Get details of a specific monitor including stats |
create_check |
Create a new monitor (returns ping URL and code snippets) |
update_check |
Update monitor settings (name, period, grace) |
delete_check |
Delete a monitor permanently |
pause_check |
Pause a monitor (no alerts while paused) |
resume_check |
Resume a paused monitor |
Debugging Tools
| Tool | Description |
|---|---|
get_check_output |
View captured stdout/stderr from recent job runs |
diagnose_check |
Analyze why a check is failing with actionable suggestions |
Example Conversations
Creating a Monitor
You: "Create a monitor for my nightly database backup that runs at 2 AM"
Assistant: Creates a monitor with 24-hour period and provides the ping URL, ready-to-use curl commands, and code snippets for output capture.
Debugging a Failure
You: "Why is my backup job failing?"
Assistant: Uses diagnose_check to analyze the monitor status, examine captured output for error patterns, and suggest specific fixes like checking credentials, file paths, or permissions.
Quick Status Check
You: "How are my cron job monitors doing?"
Assistant: Lists all monitors with their status, last ping time, and uptime percentage.
Output Capture
CronSignal can capture the stdout/stderr from your cron jobs. This lets AI assistants read the actual error messages when diagnosing failures.
To enable output capture, pipe your job's output to the ping:
# Capture output and send to CronSignal
your-backup-script 2>&1 | curl -fsS -X POST -d @- https://api.cronsignal.io/ping/YOUR_CHECK_ID
# Include exit code
your-backup-script; curl -fsS "https://api.cronsignal.io/ping/YOUR_CHECK_ID?exit_code=$?"
Environment Variables
| Variable | Required | Description |
|---|---|---|
CRONSIGNAL_API_KEY |
Yes | Your CronSignal API key (starts with cs_live_) |
Manual Installation
You can also install the MCP server globally:
npm install -g cronsignal-mcp
Or run directly with npx:
CRONSIGNAL_API_KEY=cs_live_... npx cronsignal-mcp
Links
- Get API Key - Create your API key and configure your assistant
- GitHub - MCP server source code
- Model Context Protocol - Learn more about MCP