Claude-built Mac app counts down planes visible from your yard

A Mac app built with Anthropic's Claude is drawing attention. Enter your address, rough out the outline of the trees and rooflines around your yard, and the app tracks real aircraft passing overhead in real time, telling you by voice which direction to look and how many seconds until it appears. A video posted to YouTube showed the app actually working.

The app doesn't just tell you a plane is passing by. It factors in how long sound takes to reach your ears, starting the countdown a few seconds before the plane becomes visible. In the video, the voice prompt goes something like: "Look east... in view in five, four, three."

If it's hidden behind a tree, the app says listen, not look

What stands out most is that the app folds the tree and roofline outline you drew into actual physical calculations. It only issues a "Look" prompt once an aircraft has climbed high enough to clear the trees in your yard; if a plane is hidden behind trees or a roof and can't actually be seen, the app gives only a "Listen" prompt. Every aircraft shown on screen is reported to reflect real, live air traffic data.

Built on free, publicly available flight data

Most of the data powering the app comes from public infrastructure run by volunteers. Aircraft positions pass through community aggregation sites adsb.lol and adsb.fi, which pool position signals (ADS-B) that transponder-equipped planes continuously broadcast over open radio frequencies, picked up by thousands of hobbyist antennas. Flight path data comes from adsb.lol's community route database, and ground elevation comes from Open-Meteo. Converting the address into a map and reading the prompts aloud is handled by Apple's on-device frameworks.

Here's a summary in table form:

Data type	Source
Real-time aircraft position	adsb.lol, adsb.fi (community ADS-B)
Flight path info	adsb.lol community route DB
Ground elevation	Open-Meteo
Geocoding, maps, voice	Apple on-device frameworks

Accuracy confirmed with a Rome-to-New York flight

The actual demo footage shown in the video reveals just how specific the app's calculations are. That level of detail is only possible by combining ADS-B signals with flight path data.

How it works in practice

The workflow shown in the video goes like this. First you enter your home address, then rough in the outline of the trees and rooftops around your yard. Using that terrain data, the app keeps watch on overhead air traffic, and once a plane meets the condition of clearing the tree line, it gives a voice prompt with the direction and a countdown. If a plane is blocked from view by a roof or trees, the app only issues a "Listen" prompt instead. The video didn't disclose how the app will be distributed or where it can be downloaded.

Editor's take

What this demo really shows is one thing: AI coding tools can now go beyond spitting out text or a single image, and actually build working software that stitches together multiple live public data sources and runs the physics calculations to match. In the past, a developer would have spent days digging through ADS-B API docs, writing coordinate-conversion logic, and bolting on a separate speech-synthesis API. Here, that entire pipeline came together as a single finished product.

That approach is directly useful for local teams and independent developers alike. It means that with a handful of public APIs — ADS-B, terrain elevation, mapping — you can quickly prototype your own personalized alert app. That said, this particular app is still at the demo stage with no announced distribution path, so it looks like it'll be a while before anyone can actually get their hands on it.

Expect to see more demos built the same way — coding agents paired with live data published by hobbyist communities, turned into personal alert apps. This one might be the first example of a broader shift where looking up at the sky goes from chance to calculation.