Post-purchase price adjustments
Target refunds the difference for fourteen days. Most people never find out. Receipt Snoop reads your receipts and holds on to them, so the moment you scan the lower price it can tell you what you are owed and the exact words to say at the desk.
Nothing here watches prices for you. There is no price fetcher and nothing is scraped. A lower price gets in when you scan it — which is why the scanner is the product, and why the receipts have to be in first.
Free while we are new. No card. Delete everything, for real, in one click.
The founding case
This is the case the product was built around, run end to end. The receipt, the price drop, the arithmetic and the script — none of it rewritten for the website.
TARGET
23 JUL 2026
036000291452
Balanced means the receipt agreed with itself. Every line, every discount and the tax reach the total printed at the bottom. If they had not, we would have told you that instead of guessing.
You scanned it on target.com
31 July 2026
$11.99
You scanned it. Nothing went looking on your behalf — that is the shape of this product, not a gap in it.
A price is not treated as solid the moment somebody logs it. A background worker looks for another shopper who logged the same shelf price independently: the same store, within twenty-four hours, and not you. That is only the precondition. Being in the same shop on the same day is something a fabricated receipt can arrange, so what actually counts is agreeing about the number — the same resolved product, to the cent.
| Retailer | Target |
| Where | Guest Services |
| Window | 14 days from purchase |
| Closes | 6 August 2026 |
| Days left | 6 |
| Recoverable | $18.00 |
You take this to the desk yourself. Receipt Snoop writes the words; it does not file anything on your behalf, at Target or anywhere else.
Hello — I'd like a price adjustment. I bought BLUEY FIGURE 4PK (item 036000291452) here on 23 July 2026 for $29.99. It's now $11.99 on their website, which I checked on 31 July 2026. Your price-adjustment policy runs 14 days from purchase, so this is within the window. I'd like the difference of $18.00 refunded to my original payment method.
Policy read 1 August 2026 from target.com/help/article/000062256, by automated research against that page. No human has re-read it since.
Same overpay, different evidence
The arithmetic was certain both times. The entitlement was not. A claim is never worded more firmly than the evidence behind it, because you are the one who has to stand at the service desk.
| Savings | $18.00 |
| State | advisory |
| Confidence | worth asking |
| Recoverable | $0.00 |
Hello — I wanted to ask about a price adjustment. I bought BLUEY FIGURE 4PK (item 036000291452) here on 23 July 2026 for $29.99. It's now $11.99 on their website, as of 31 July 2026 — a difference of $18.00. Do you offer a price adjustment for purchases within a certain window? If so, I'd like to request one.
Advisory money is left out of the headline figure on purpose. A total you cannot go and collect is a lie with a decimal point in it.
| Savings | $18.00 |
| State | actionable |
| Confidence | policy confirmed |
| Recoverable | $18.00 |
| Days left | 6 |
| Venue | Guest Services |
The only thing that changed is whether we had grounds to promise anything. The wording tracks that exactly.
Every figure says where it came from
A receipt that does not balance was either misread or edited, and we say which we think it was rather than quietly using the numbers anyway.
A figure with no badge beside it is a bug in this product, not a design choice.
Getting your receipts in
You get your own address at rcpt.receiptsnoop.com. Forward a retailer's receipt to it, or set a rule and forget about it. When the signature on the mail checks out, that receipt is the strongest evidence in the system — the retailer sent it, and nothing was altered on the way here.
Highest trust. No account passwords, ever.
Point your phone at the till roll. There is no single image treatment that reads every receipt, so it tries the cheapest one first and lets the receipt's own arithmetic say whether it worked — up to three attempts, stopping at the first one whose sums close. A photograph we cannot read is refused with advice, and kept anyway. It is all that is left of a receipt that is already in a bin.
The in-store one, and the only way a lower price ever enters by itself. Scan the barcode off the shelf tag and log what it costs today. On the web there is no camera worth relying on, so it degrades to typing the digits printed under the bars — eight, twelve or thirteen of them, the same barcode either way. Tick anything the tag says: clearance, doorbuster, members-only. Those are the conditions retailers use to refuse a claim, and we would rather know now.
So: two ways a receipt gets in, and this one way a price does. There is no type-it-all-in-by-hand screen yet — the API accepts one, the app has no route to it.
No bank login. No retailer password required. Email receipts prove themselves, which is what lets this product stay out of your accounts.
Coverage, stated plainly
Thirty-eight retailers are in the registry and all thirty-eight now have a policy record. Thirty-two of those were read against the retailer's own published page inside the last ninety days, which is how long a record stays trusted before the app starts marking claims advisory again. Sixteen of the thirty-two adjust a price after you have paid. Fifteen of those publish a deadline you can count down; one adjusts and publishes no deadline at all. Here is the whole picture, including the parts that are not finished — a coverage map you cannot check is not a coverage map.
| Retailer | Window | Where | Notes |
|---|---|---|---|
| Costco | 30 days | Online | Online orders only. A warehouse purchase has no published window |
| Fred Meyer | 30 days | Customer Service Desk | Home, Apparel and Home Electronics only. Groceries do not qualify |
| Home Depot | 30 days | 1-800-430-3376 | Will adjust against a competitor it matches, not only its own price |
| Wegmans | 30 days | Service Desk | Home & Entertaining only. A drop on food gets nothing |
| Best Buy | 15 days | Customer Service desk | 60 days with My Best Buy Plus or Total |
| ALDI | 14 days | The store manager | ALDI Finds only, at the store you bought from |
| JCPenney | 14 days | Customer Service | Will match any retailer's advertised price |
| Kohl's | 14 days | Customer Service | Its own lower price, original receipt |
| Menards | 14 days | Any Menards cashier | Only while the sale is still running, which is usually the shorter clock |
| REI | 14 days | Online form | Its own price. Outlet items do not qualify |
| Staples | 14 days | Customer Service | Same item cheaper in store or on Staples.com |
| Target | 14 days | Guest Services | Target.com and the store you bought from |
| Macy's | 10 days | Online form | You have to file it yourself — Macy's insists |
| Meijer | 10 days | Service Desk | Only when the item lands in the Meijer weekly ad |
| Nordstrom | 10 days | Online form | From the date it shipped, not the date you ordered |
Every window above was taken off the retailer's own published policy page, field by field, with that page's URL and the date it was read stored next to the record. When one goes ninety days without a re-check, it stops being treated as confirmed and its claims go back to being questions.
Four of these are narrower than a window alone suggests, and the claim says so rather than the table hiding it: ALDI's covers ALDI Finds, Wegmans' covers Home & Entertaining, Fred Meyer's covers Home, Apparel and Home Electronics, and Menards' runs only while the sale flyer that triggered it is still on. A weekly shop at any of the first three qualifies for nothing at all.
What "verified" means here, precisely: automated research read the retailer's own published page and recorded the URL, the date and the exact sentences it relied on. A person has not sat down and re-read every one of them. That is a real difference and it is the reason a wrong window is treated as a bug worth an email rather than a disclaimer.
So there is no countdown to show you, and inventing one would be worse than useless — an earlier draft borrowed the 90-day return window from the same page, which would have told somebody at day 85 that they had five days left and somebody at day 100 that they were out of time. Both invented. The overpay is surfaced with the money on it, the script asks rather than asserts, no date is quoted, and the advice is the only honest one available: ask promptly, because an unpublished window is not a long one, it is an unknown one. It also means the claim can never expire — you cannot miss a deadline nobody published. Lowe's will only take the request from you, so this is one you make yourself, by phone or chat.
So we do not send you to the desk. What we do send you is the other move: return it and buy it again at the lower price. That advice needs a return window we have actually read, and all four of these now have one on file — so an overpay at any of them is raised as advisory, with the money, the reason there is no claim, and — for as long as that window is open — the days left on it instead of on an adjustment that does not exist. Once it closes there is genuinely nothing to say, and the feed stays quiet. Recording these as "unknown" would waste your afternoon; recording them as "does not price match" would be false.
One retailer in the corpus sits here, and it is the only one that reaches this branch of the code at all. It is worth the row because it is the shape a shopper most often mistakes for a refusal — a till-side guarantee is not a promise about tomorrow, and treating it as one sends somebody back for something that was never offered.
An overpay at any of these shows up as advisory. You get the number and a question to ask — not a promise and not a deadline on something nobody has published. The return-and-rebuy advice does not fire here either, and that is deliberate rather than an omission: it needs a recorded refusal to adjust, and these eight have not refused. They have said nothing. Telling you to return something on the strength of silence would be guessing at their policy on their behalf. Note also that the four seven-day windows cover delivery and pickup orders only; walking into the shop with a receipt is not what that number describes.
Six retailers refuse automated clients outright, so nothing about them is recorded in either direction. An overpay at any of them is still detected and the arithmetic is still exact; the claim is worded as a question and the money stays out of your headline total.
Getting a window wrong does not produce a slightly wrong number. It sends a real person to a service desk for nothing. That is why this list grows one retailer at a time, each one taken from its own published terms and dated, rather than at the speed of a press release.
The refusals
There is no price fetcher and nothing is scraped. A lower price gets in when you scan or type one you saw, when another shopper who opted into sharing scanned it, or when it appears in a weekly ad somebody imported by hand. There is no fourth way in. What your receipts buy you is that the moment one of those happens, the answer is already worked out.
If a retailer's "you saved $138" cannot be interpreted with certainty, no original price is worked backwards from it. A guessed market price is worse than none: it contaminates every comparison after it.
A photographed receipt is Captured and stays Captured, whatever the upload claims about itself. Nothing a client sends can grant it trust.
Where there is an assistant, it is handed figures we already worked out and asked to write English. Any number in its answer that was not in the facts is rejected outright. Extraction, the price arithmetic and claim windows never touch a model at all.
When the solver and the receipt disagree, you see both, and the printed one is the one we keep.
This week's ads
Today we hold no circulars at all. When one is in, it is laid out in page order with each item linking out to the retailer's own page, and it does the part only a product holding both halves can do: an item you have bought before is marked, with what you last paid sitting next to what it costs this week. "$9.99 this week, you last paid $19.99."
Thirty-eight retailers are in the registry. Twenty-seven publish a weekly ad. Exactly one of those — Best Buy — can be read programmatically today. There is no public weekly-ad feed in the United States, and most of the alternatives are somebody else's terms of service. So the first ads go in by hand, and until they do the screen says it is empty and why, rather than looking broken. Every ad we hold says the date it was imported.
"With your Card" is not your price if you do not hold the card. "2 for $5" is not the price of one. Those conditions are stored, shown on the ad with the condition named, and never turned into a claim. When an ad's week ends, its prices are retracted rather than left lying around.
Whose side this is on
There is an export too: every receipt, every price you logged, and every intake attempt including the ones that failed — because "you sent us this and we could not read it" is a fact about your data as well. And we can answer a question most companies cannot: who at Receipt Snoop has looked at your receipts. Staff reads are recorded, opening a basket needs a typed reason, and that log cannot be edited or deleted by anyone, including us.
Straight about the stage
No testimonials on this page, no counter of money saved, no logos. There is nothing to put there yet, and inventing it would be a strange way to open a relationship with someone whose receipts we are asking for.
What does exist: the arithmetic, tested to the cent. Thirty-two retailers whose policies have been read against their own published terms and dated, sixteen of them adjusting after purchase, fifteen of those firm enough to put a date on. A photograph pipeline that admits when it cannot read something. No weekly ads yet — the first ones go in by hand. And a coverage list that gets longer one verified policy at a time.
If a window is wrong, tell us and we will fix the record. A wrong policy is a bug here, not a disclaimer.
It takes a minute. If it finds nothing, it says nothing — the feed stays quiet rather than inventing something for you to do.