What is FeedBunk AI?

Most dairies already have a trail camera pointed at the feed bunk. The photos pile up on an SD card, and then one of two things happens: somebody spends hours scrolling through them marking down what they see, or nobody looks at them at all.

FeedBunk AI reads those photos for you. You upload a card’s worth of images, and you get back a report showing every feed delivery, every push-up, and every stretch where the bunk sat empty — with the times attached, around the clock, including the overnight hours nobody is there to watch.

What you need

Just three things, and you probably have all of them already:

  • A trail camera pointed down the feed alley. Any brand with a time-lapse mode works — see Camera requirements and settings.
  • An SD card you can pull and carry to a computer.
  • A computer with an internet connection, to send the photos to us.

No streaming cameras, no sensors, no wiring, no monthly hardware. If your camera can take a picture every ten seconds and stamp the time on it, that is enough.

What you get back

Your report lands on one page and covers four things:

  • Empty bunk time — how many hours the bunk sat empty, the longest single stretch, when it usually happens, and how much of it falls overnight.
  • Feed deliveries — how many per day, what times they land on, and how tightly they hold to that schedule day to day.
  • Push-ups — how many per day, how long the gaps between them run, and whether each delivery got a push-up within two hours.
  • A day-by-day timeline — every event on a 24-hour strip per day, so you can see the shape of a week at a glance.

Every number is backed by the actual photos. Click any event and you see the frames it came from, including the five frames before it started, so you can check the call yourself.

How the work flows

  1. Set the camera to time-lapse and let it run for one to two weeks.
  2. Pull the SD card and put it in your computer.
  3. Send us the photos — fastest with the FeedBunk AI Compressor desktop app, or straight from your browser.
  4. Mark the corners of your bunk once, so we measure the right strip of concrete.
  5. Read your report. Correct anything that looks wrong, and the numbers update.

The whole loop is walked through step by step in Quick start: from camera to report.

What it does not do

FeedBunk AI is not a live monitor. It works on photos after the fact, not on a streaming feed, so it tells you what happened last week rather than what is happening right now. That is the trade — you get full 24-hour coverage of a real two-week stretch, from a camera that costs a fraction of a connected system.

It also does not weigh feed. It reads what the camera can see: whether feed is there, when it arrived, and when it got pushed back within reach.