Ninja air fryer, 4qt
9486%Target · 3 stores near you
▲ 412 · 38 confirmed · 2h
A repost tells you a Target markdown exists. It does not tell you which store, which aisle, or that the endcap at Mueller is already empty. Goatavo watches the source itself — store inventory endpoints, register scans, drop calendars, portal rates, airline seat maps — and puts it on your lock screen in 4.2 seconds, with the aisle already in the notification body.
Ryobi 18V drill/driver kit rang up at $0.01
100%Home Depot · Slaughter Ln · aisle 14 · 3 scans in the last 6 min
4.1s from the register
Ninja air fryer 4qt dropped to $17.98
86%Target · Mueller · aisle D14 · 4 on the shelf, was 9 at 6:02a
matched 2 of your rules
2 ANA business seats opened, JFK → NRT
91Oct 11 · 60,000 points round trip · seat map polled 4 min ago
9 more in Alerts · 2 batched for 9:00a
Ryobi 18V drill/driver kit rang up at $0.01
100%Home Depot · Slaughter Ln · aisle 14 · 3 scans in the last 6 min
Ninja air fryer 4qt dropped to $17.98
86%Target · Mueller · aisle D14 · 4 on the shelf, was 9 at 6:02a
2 ANA business seats opened, JFK → NRT
91Oct 11 · 60,000 points round trip · seat map polled 4 min ago
Air fryer count flipped 0 → 4, price $129.00 → $17.98
Target inventory endpoint · store 1234
Every alert product claims to be fast. Here is the entire chain for one real-shaped event — a Ninja air fryer going from $129.00 to $17.98 at Target Mueller, store 1234 — measured on both routes.
Target's own inventory endpoint flips store 1234 from 0 to 4 and the price from $129.00 to $17.98.
The diff lands in the change queue with price, on-hand count and aisle D14 already attached.
86% off, all-time low, 1.2mi from you. Matches 2 of your 7 rules and clears both guards.
Push delivered: $17.98, 4 left, aisle D14. Nothing to open, nothing to search for.
Nobody is watching the endpoint. The four units sit on the endcap in aisle D14.
A shopper photographs the shelf tag and posts it to a Facebook group with no store name.
Two deal accounts copy it. One adds an affiliate link to the full-price online listing.
The aggregator's digest goes out. No store, no count, no aisle — 'check your Target.'
Mueller sold its four at 7:41, 7:44, 7:52 and 8:03. You drove to a repost of an empty endcap.
The gap between the two rails is 25 minutes and 55.8 seconds — longer than the 25 minutes four air fryers lasted on that endcap, in a store 1.2 miles from you.
Get the 4.2-second pathEvery alert service you have used made you learn a different form for every kind of thing. Goatavo has five fields — watch, where, when, only if, then — and they do not change when the subject does. Below is the same builder filled in five ways. Read the labels, not the values: they are identical.
A penny item. A portal rate. A gift card discount. A transfer bonus. A bank’s direct-deposit window. A search you saved at Target Mueller. A URL we have never seen before.
Same five fields. If it changes and we can see it, you can put a rule on it — and goat will write the first draft of the rule if you describe it in a sentence.
Nobody can evaluate an alert product that has not fired yet. So the first thing Goatavo does is run your rule backwards over real events and show you, hour by hour, what you would have been told and whether the thing was still there when you got there.
watch clearance, any category · where 6 stores inside 12mi of 78702 · when it crosses 70% off · only if 2+ still on the shelf · then instant push
You would have seen all 6 while stock lasted.
Median 30 min of runway between the push and the last unit leaving the shelf. Shortest was 14 min, at Walmart Ben White.
You would have seen 5 of 6, after they sold out.
The sixth never made the digest at all — it was collapsed into a line that read “12 more clearance items.” Latest delivery was 12h 16m after the boots were gone.
Target · Mueller · aisle D14
4 on the shelf · last one left at 8:03a
25 min of runway before the last one left.
57 min after the last unit left.
Target · Anderson Ln · aisle G03
6 on the shelf · last one left at 5:47p
95 min of runway before the last one left.
13 min after the last box.
Home Depot · Slaughter Ln · aisle 14
3 on the shelf · last one left at 7:19a
77 min of runway before the last one left.
1h 41m after the last kit.
Walmart · Ben White · aisle H11
5 on the shelf · last one left at 11:58a
14 min of runway before the last one left.
6h 2m after the last set.
Dillard's · clearance center · sizes 7–10
2 left in the size run · last one left at 8:44p
23 min of runway before the last one left.
the 6:00p digest had already gone — 12h 16m late.
Home Depot · Sunset Valley · seasonal reset
2 on the pallet · last one left at 6:31a
34 min of runway before the last one left.
collapsed into '12 more clearance items'.
$209.45 against $1,265.99 of sticker, across six events in one week, on one rule, in one zip code. Nothing here required a faster thumb — only a shorter gap between the shelf changing and you knowing.
The dry run is free, it runs before you have an account, and it uses last week’s real events at the stores nearest you. If your area is quiet, it will say so in plain language rather than inventing six deals.
A notification that says 'deal alert!' is worth nothing. Every Goatavo push carries what changed, where, how many are left, which source saw it, and how long ago — because those five facts are the entire difference between acting and scrolling.
Ninja 4qt air fryer went $129.00 → $17.98 at Target Mueller.
Walmart Ben White went 0 → 5 on the Dyson V-series filter set.
3 left at Home Depot Slaughter Ln, down from 7 at 5:40a.
SB Dunk Low draw opens 10:00 ET — pushed at 9:50, size run attached.
Sapphire Preferred at 150k — first time above 100k since March.
2 ANA business seats, JFK → NRT, Oct 11, 60k round trip.
34 items live on this week's Dollar General Tuesday reset.
TopCashback at 15% for Walmart — 6 points above Rakuten, today only.
All eight are on the free tier. What Pro moves is the delivery — how many go instant, and how often we poll the slow sources.
See what Pro changesTarget · 3 stores near you
▲ 412 · 38 confirmed · 2h
ANA · 4 dates open in October
▲ 88 · 1d
TopCashback · beat Rakuten by 6%
▲ 264 · 26 confirmed · 45m
Tapping an alert does not open a browser tab. It opens the card — with the aisle, the on-hand count, who confirmed it and when, the 90-day price line, the resale comps if you want them, and the card that earns the most on this purchase.
Home Depot · penny list · scan to verify
▲ 1.4k · 96 confirmed · 12m
Sapphire Preferred · ends in 6 days
▲ 288 · 12 approved · 5h
Everybody who has run a cook group knows the failure mode: the alerts are fast, then they are constant, then they are wallpaper, then you turn them off and miss the one that mattered. Four controls exist specifically to stop that, and all four are on by default.
Anything under the bar you set collects into one digest at 9:00a and 6:00p. Instant is reserved for the rules you actually marked 'I would run for this' — most people have two or three.
Digest at 09:00 and 18:00 · 3 rules instant
22:00 to 06:30, nothing makes a sound. The exceptions are the rules you tag Override — usually one drop and one mistake fare. Everything else waits for the 9:00a digest and says how long it waited.
22:00–06:30 · 2 rules can override
A community shelf claim waits for a second independent scan before it pushes. Endpoint-sourced events don't have to wait, and the notification says which kind it was, every time.
Community events need 2 scans · endpoint 1
The same air fryer at three Targets is one notification with three store rows. The same penny item reposted forty times is one line. If a rule fires twice inside 20 minutes, the second one edits the first instead of buzzing again.
20-min collapse window · 3 stores in one card
What the median beta tester was getting before they switched. Roughly one every three waking minutes, most of them reposts of each other.
Median across the same people, eight weeks in. 2.1 of them arrived instant; the rest waited for 9:00a or 6:00p.
of the alerts people muted in beta were community reposts of something they had already been told. Dedupe catches those before they buzz — it is most of the distance between the 40+ a day an unguarded rule produces and the 11 above.
Free is not a demo. Every alert type, every field in the builder, the aisle, the counts and the confirmations are free forever. What Pro buys is delivery: how many rules fire instantly instead of waiting for a digest, and how hard we poll the sources that move slowly.
Three rules and five instant pushes a day is genuinely enough for one product and one store. It is not enough for a drop calendar, and we are not going to pretend otherwise on the pricing page.
When a free rule gets batched, the notification says so: “held 3h 12m — 2 of 4 left when it was found.” You always know what the cap cost you, in units.
Founding members keep the launch price for as long as the account stays open, and get the 90-day dry run on day one.
“I was in four Facebook groups, two Discords and a text thread. Everything I actually used is one feed now, and the aisle number is right.”
“The comps are on the card. I stopped opening a second app to work out whether a $17 air fryer was worth the drive.”
“It knew my 5/24 dates and told me to hold. The 150k landed nine days later. That's the whole product.”
“Nine minutes of warning is the difference between a pair and a screenshot. I cancelled three cook groups.”
“The seat alert polls the airline, not a blog. I got two ANA business seats on a Tuesday afternoon.”
“The list used to be a blog post with 400 comments. Now it's scannable and mapped to my four stores.”
A restock and a bank bonus have nothing in common except that both are worth knowing about first. Each row below is the rule people actually write on that page, and how hard we poll for it.
The alert is the same object whether it wakes your phone, sits in the Alerts tab, or arrives in tomorrow morning's digest. Only the delivery changes.
Only the rules you set to instant. Aisle and on-hand count on the first line, tier badge on the right, source named underneath.
Everything the engine matched, newest first, with the ones you have not opened marked and a mute button on every row.
One card per rule, deduped across stores, each stamped with how long it sat. The digest never pretends it arrived on time.
Median 4.2 seconds from the source event to the push leaving our servers, measured across the beta. The slowest tenth lands inside 11 seconds. After that it is Apple's and Google's push network, which is usually under a second and which we do not control — so the notification carries the detection timestamp and you can see the whole chain yourself.
Every alert names its source: retailer endpoint, register scan, brand calendar, airline seat map, portal crawler, or community confirmation. If an alert came from a repost we say so and it is never marked instant. That label is the entire promise.
It happens — endpoints lag resets and cycle counts. That is what the confirmation count is for. An endpoint claim with no scan behind it says so on the card, and if two people scan it and get a different price the alert is retracted and everyone who got it is told.
Only if you tag a rule Override, and you have to do that per rule. Quiet hours default to 22:00–06:30. A held alert arrives in the morning with the honest line: held 7h 12m, sold out at 04:41.
Yes. The builder is one shape — watch, where, when, only if, then — and the 'watch' field takes a product, a store, a brand, a search, a route, an offer page or a URL. If it changes and we can see it, you can put a rule on it.
No. It runs before you have an account and before you have paid for anything, on the last 7 days of real events. That is deliberate — it is the only honest way to sell an alert product that hasn't fired for you yet.
Push and email at launch, with a per-rule choice. SMS is on the list for the drop and mistake-fare crowd, where the phone is in a pocket and the window is under ten minutes.
In beta the median hunter got 11 a day, 2.1 of them instant. The people who got 40+ had written a rule with no guard on it — the builder now warns you before you save one, with the count it would have fired last week.
Want to see the engine pointed at drops instead? The drop calendar page walks through T−10.
Join the launch list and your first email is a dry run on your own zip code — last week's real events, and exactly what you would have caught while stock lasted.