Petlibro has launched the Granary 2 series, a new lineup of automatic dry-food feeders that put a scale under the bowl and, on the pricier models, an AI camera above it. The pitch: stop guessing whether your cat actually ate and start seeing real intake data. TechCrunch AI tested the mid-range Granary 2 Vision and called it one of the best feeders they’ve tried.
Here’s what stands out about the launch, and where the catches are.
1. Four models, one core upgrade
The series runs from $129.99 to $249.99, according to TechCrunch AI:
- Granary 2 ($129.99): App-controlled feeding, precise portioning, intake tracking.
- Granary 2 Vision ($189.99): Adds an AI camera that recognizes up to 10 cats and logs each one’s eating habits.
- Granary 2 Vision Duo ($199.99): Two dispensing chutes and two bowls, so each pet gets its own portion.
- Granary 2 X ($249.99): Uses AI recognition for individualized portioning, aimed at pets with specific dietary needs.
The shared upgrade across all four is a built-in scale. It weighs food in the tray, so the feeder knows both what it dispensed and what’s still sitting there. That’s the difference between “a meal was released at 7 p.m.” and “your cat ate 80% of it in four minutes.”
2. The data is the product
The feeder tracks intake, eating duration, feeding frequency, favorite eating times, and eating speed. The companion app rolls that into daily, weekly, monthly, and yearly views.
Why this matters: Cats hide illness well, and appetite changes are often the first visible signal. A feeder that builds a baseline of normal behavior can flag a drift you’d never catch by eyeballing the bowl. TechCrunch AI frames this as the feature that makes the Granary 2 worth considering for owners who closely watch weight or portion sizes.
3. Two feeding modes plus a manual button
- Smart Refill Mode: Keeps the bowl topped up with small, fresh portions throughout the day and stops once your daily limit is hit. Built for grazers.
- Scheduled Mode: Classic set times and set portions.
- Manual button: For off-schedule feeding when you want it.
4. The AI camera does the sorting in multi-cat homes
On the Vision model TechCrunch AI tested, the 1080p camera identifies which cat is at the bowl and assigns the meal to the right profile automatically. It only opens when it recognizes the correct pet, which solves the classic problem of one cat eating the other’s diet food.
The camera also brings night vision, two-way audio, and a recorded call-to-dinner message that plays from the feeder itself. App alerts cover low battery, low food, and jams, and a rechargeable backup battery covers power outages.
5. The subscription catch
This is where I’d slow down before buying. Some of the advanced health tracking and monitoring requires a Petlibro Care subscription starting at $59.99 per year. Cloud recording and playback on camera models needs a separate Video Cloud plan at $119.99 per year, as detailed in TechCrunch AI.
So the $189.99 Vision can quietly become a $370 first-year purchase if you want everything it advertises. That’s the pattern across pet tech right now: hardware margins are thin, so companies push the intelligence layer into recurring revenue. Petlibro isn’t unusual here, but the fine print matters.
6. The cheaper alternative
If you want the scale without the AI, TechCrunch AI points to Pawsync’s $79.99 feeder as the closest comparison. It also weighs portions and includes settings to prevent under- and overfeeding, minus the camera, the pet recognition, and the subscription stack.
7. Why this matters beyond cats
Smart feeders used to be timers with a hopper. The Granary 2 shows how quickly consumer hardware is absorbing on-device recognition and behavioral analytics, the same playbook we’ve seen in smart doorbells and sleep trackers. Identifying 10 individual cats by camera and gating a physical mechanism on that recognition is a real computer vision task running in a $190 kitchen appliance.
The open question is whether owners will pay yearly for insights about their cat’s eating speed. Petlibro is betting they will. The full hands-on review at TechCrunch AI has more detail on how the Vision held up in testing.