StatHat MCP. Log every metric—from counters to complex values.

Forget writing boilerplate SDK code or manually updating dashboards just to log a metric. When you connect your AI client to StatHat, you're giving your agent an instant pipeline for performance data. It’s built for developers who need real-time tracking—you don't want to leave the chat window every time something happens, so StatHat lets your agent post custom metrics and stats directly to your observability stack using four specific tools.

When you use this server, your AI client acts as the bridge. Instead of telling your agent what to do with code, you just tell it what happened; then, the right tool handles sending that data instantly. You'll find the functionality covers everything from logging a single event count to dumping massive sets of diverse metrics all at once.

Increment Event Counters: You can use post_classic_counter when you need to track simple, discrete events. If something happens—say, a user clicks a specific button or an API call succeeds—you instruct your agent to run this tool, and it increases the count for that specific event key by exactly one. This is how you measure total occurrences of any defined action.

Record Specific Numerical Values: For continuous numbers, you'll use post_classic_value. Think latency measurements in milliseconds, CPU usage percentages, or resource availability; these are all values. When you call this tool, your agent sends a precise number to a dedicated stat key, allowing you to track rates and magnitudes over time.

Create Metrics on the Fly: Sometimes you don't know if a metric key exists yet—you just need to log it immediately. That’s where post_ez_stat comes in handy. You give this tool a descriptive name, and StatHat automatically handles creating the metric for you, logging that new stat without any setup required on your end.

Batch Log Multiple Statistics: When you're dealing with large amounts of data—maybe counts and values from several different sources at once—you don't want to make multiple API calls. You use post_json_stats. This tool accepts structured JSON data, letting you send diverse metrics (counts, specific numerical values, and timestamps) in a single, efficient call.

It’s how you keep your logging fast when you're tracking everything.

If you need to track the number of times a user hit 'submit,' you invoke post_classic_counter. If you're measuring the average time it took for a database query to run, you use post_classic_value with that continuous time. Need to log a new metric called 'Payment Gateway Failures'? Use post_ez_stat just by giving it the name.

And when your application generates dozens of data points—like counts from three different endpoints and values for two different service latencies—you bundle all that crap into one JSON payload using post_json_stats. This server gives you granular, immediate control over metrics logging without forcing you to write a single line of extra code.