Hack Your Taps - Again!

May 2, 2026

Hack Your Taps – Again!

Back in 2020 we told you how to hack your taps to build custom integrations on top of your Kegtron data. The tools have gotten a whole lot more interesting since then. Time for an update.

Whether you want a custom dashboard and have never opened a code editor in your life, want instant push notifications when a keg goes low, or are running Node.js in your sleep and want live pour data at 200ms intervals, we have something for you. Three paths, three different audiences.

Path 1: No Coding Skills? Meet Claude.

You cannot hide from AI hype these days, but when it comes to software the hype is justified. AI assistants, most notably Claude, are now good enough to create working software integrations from plain English descriptions. You do not need to understand any of it. You just need to know what you want.

We put together a step-by-step guide called the Pro AI Integration Guide that walks you through the whole process. The short version:

Grab your API key from the Account tab in your Kegtron Cloud Dashboard. (PSA: never share your key or post online!)
Open Claude (or another AI assistant – we think Claude gives us the best results).
Paste a link to the REST API spec and describe what you want in plain English.
Run what it gives you. The guide includes a ready-to-use example prompt that produces a working keg level dashboard in a single conversation.

The AI writes the code, and the guide walks you through running it. Black magic!

10mins with Claude

What Could You Actually Build?

Our Cloud Dashboard already does a lot of cool stuff already, no need to reinvent the wheel. That said, there are a lot of cool things you could do to make things your own:

A custom tap list display with your branding, colors or background image
An activity tracker heatmap showing when the action is happening
A Slack alert when any keg drops below a threshold
Customized SMS messages based on different actions (Friday night party alert!)

The AI can build all of these. And when something doesn’t work, you just describe the problem and it fixes it. That back-and-forth is completely normal, even for experienced developers.

Claude Code: The Power-Up Version

If you want to take it further, the guide also covers Claude Code, available in the Claude desktop app or CLI. Instead of copying code out of a chat window and saving it manually, Claude Code writes files directly to your computer, runs them, catches its own errors, and tells you when it is done. It removes a lot of the friction once you are building something with more than one file.

For most people, the right move is: start in the browser with claude.ai, get something working, then graduate to Claude Code if you want to keep building on it.

Read the full AI Integration Guide

Path 2: Want Instant Alerts? Use WebSockets.

The REST API works great when you are happy fetching data on a schedule. But polling has a cost: you are always a few seconds behind. It’s also not very efficient to have frequent, nonstop polling. The kind of brute force approach won’t win many coding style points.

WebSockets flip that around. Instead of your app asking Kegtron for updates, Kegtron pushes a notification to your app the instant anything changes. A pour completes, a keg goes offline, a level drops below your threshold: your code knows immediately.

Why Choose This Over REST

If any of these sound familiar, WebSockets are probably the right call:

You want a Slack message or SMS the moment a keg hits a low threshold, not a minute later
You are building a live dashboard and do not want to hammer the REST API every few seconds
You need to trigger something in another system as soon as a pour is recorded

The REST API is pull. WebSockets are push. For event-driven stuff, push almost always wins.

Level of Effort

This one sits comfortably between REST and UDP in terms of complexity. You need to be comfortable writing a bit of JavaScript (or use an AI assistant to do it for you), but there is nothing exotic here. A working listener is around 20 lines of Node.js. Open a connection to the WebSocket endpoint with your device key, listen for message events, parse the JSON, and act on whatever you care about.

The main thing to plan for is reconnection. WebSocket connections can drop, so your code needs to handle that gracefully and reconnect with a short delay. The API docs include a ready-to-run Node.js example that covers the basics including reconnect handling.

One connection covers one device. If you are monitoring multiple keg monitors, open one connection per device and fan the events into whatever processing logic you need.

Path 3: Serious Skills? Meet UDP Streaming.

The REST API ad/or WebSockets are great for 95% of use cases: dashboards, reports, integrations with other systems. But there is a limitation: there is a 1 to 30 second delay between the end of a pour and when that data is available through the cloud. If you need it quicker brute force polling is not the answer.

For anything where you need to react during a pour, not after it, that latency is a problem.

That is what UDP Streaming is for.

How It Works

When enabled, your Kegtron Pro keg monitors broadcast live JSON packets over your local LAN using UDP multicast. During an active pour, packets go out every 200 milliseconds. Between pours, they drop back to once per second per port.

The key field is curServ: the accumulated volume of the current serving in milliliters, in real time. This value resets automatically when a pour ends. Combined with the rest of the device payload (beverage name, keg volume, port state), you have everything you need to build live displays that update as the pour is happening.

A few things that make UDP multicast particularly useful for in-venue setups:

- Device discovery required. Every receiver just subscribes to the same multicast group address and port. No IP addresses to manage, no DNS, no pairing.
- Multiple senders, multiple receivers. All your keg monitors stream to the same group. All your receivers get everything. Use the serialNum field to sort packets by device.
- Sequence numbers included. The num field lets receivers detect out-of-order or dropped packets if that matters for your application.
What Is It Good For?
• Animated fill-level displays behind the bar, the kind that show a glass filling up in real time as a pour happens (real world UDP-powered display courtesy of CleverPub)
• Automated filling machines or flow control systems that need to act on volume in real time
• Any custom in-venue display or device on the same local network that needs sub-second response to pour activity
UDP Streaming is disabled by default. Enable it per device from the Settings page in your Cloud Dashboard, or via the Kegtron API.To test it, grab Packet Sender (free, cross-platform), join the multicast group, and watch the JSON payloads roll in. Once you can see the packets, the rest is up to you.Full technical details on packet timing, payload schema, RSSI recommendations, and setup steps are all in the UDP Streaming Guide.Read the full UDP Streaming Guide
The Full Picture
Not gonna lie: Kegtron Pro has the most comprehensive, open software interface of any keg monitoring system on the market. Whatever you want to create, our three API options give you the tools to build it. To recap:
- REST API: Ask a question, get an answer. Best for dashboards, reports, and integrations with other software.
- WebSocket: A persistent connection that pushes notifications the instant something changes. Best for live pour alerts and real-time displays that update as events happen.
- UDP Streaming: Local-network multicast at 200ms intervals during a pour. Best for in-venue displays and anything that needs to react mid-pour.
If you are just getting started, the REST API plus an AI assistant will get you further than you might expect. If you want event-driven reactions without polling, WebSockets are a short step up. And if you need to react mid-pour on the local network, UDP Streaming is in a class of its own.
What will you build? Tag us on Instagram @kegtron_team and let us know!