Psst: A Shared Shopping List That Remembers Prices — Scan Receipts to Compare Automatically

Psst is a shared shopping list app that scans receipts and tracks unit prices to help you spot price changes.
Psst is a shared shopping list app for iOS and Android that scans receipts, extracts each line item, and records the last purchase price and store location to help users track price changes. Its design philosophy is deliberately restrained: it stays silent when recognition is uncertain rather than forcing a conclusion; comparisons are based on unit price rather than total amount to avoid misleading results from different purchase quantities; and prices with different units of measurement are never directly compared. Each scanned entry also includes a cropped image of the corresponding receipt row for manual verification. This "say less, never mislead" product logic is rare in today's feature-heavy app ecosystem, with its differentiation built on data reliability rather than feature breadth.
A Shopping List That Solves "How Much Did I Pay Last Time?"
One of the easiest things to forget while shopping is what you paid for something last time. Psst is a shared shopping list app available on iOS and Android, and its core selling point isn't collaboration — it's remembering the price you paid for each item. Just snap a photo of a receipt, and the app extracts every line item in a structured format, showing you what you paid last time and where you bought it.
This kind of tool sounds simple, but what's genuinely hard to get right is recognition accuracy and "honesty." Psst launched on Product Hunt with 94 upvotes and 9 comments, ranking #10 for the day, and was categorized under Android, Productivity, Task Management, and Notes.

Turning Receipts into Structured Data
Psst's entire workflow revolves around receipts: after you take a photo, the app extracts each line item individually and attaches a cropped image of the corresponding row from the original receipt. This detail is surprisingly practical — when you're unsure whether the recognition was accurate, you can check it against that specific line on the original receipt without hunting down the whole slip.
Each entry also shows the last purchase price and which store it was bought from. For anyone who regularly compares prices across multiple supermarkets, this means no more relying on memory to figure out whether something got more expensive or which store is cheaper.
The underlying technology is OCR (Optical Character Recognition), which converts printed or handwritten text into structured, editable data through image analysis. However, shopping receipts are a particularly tricky scenario for OCR: small fonts, crumpled paper, fading thermal print, product names full of abbreviations or internal codes, and prices, quantities, and units all packed tightly together. Many competing apps display the raw OCR output after scanning, resulting in high error rates. Psst's "row-level crop" design is essentially a visual fallback mechanism for OCR uncertainty — when the algorithm isn't confident, users can quickly cross-check the original receipt with their own eyes, without having to dig out the physical slip.
Restrained by Design: When in Doubt, Say Nothing
What's most interesting about this product is the developer's philosophy. According to the official description, "Most of the work went into not saying so when it isn't sure." This cuts right to the heart of a common flaw in OCR and price-comparison apps: in an attempt to seem intelligent, many tools force a conclusion and end up misleading users instead.
Psst takes the opposite approach — say less, not more. This restraint shows up clearly in two specific rules.
Prices with Different Units Are Never Compared Directly
If two prices use different units of measurement, Psst won't compare them side by side. This prevents apples-to-oranges mistakes and ensures that any conclusions drawn are actually reliable.
Mixed units are especially common with fresh produce and bulk goods: the same type of nut might have been purchased by weight last time and by bag this time; the same brand of milk might appear as 250ml in one receipt and 1L in another. If you simply subtract the two price figures or calculate a percentage change, the result is not only meaningless but creates a false signal of price movement. Many budgeting and price-comparison apps quietly ignore unit differences to display a "smart" up or down arrow in their UI — and users rarely notice the underlying error. Psst explicitly refuses cross-unit comparisons, preferring to leave the field blank rather than surface a number that could lead to a bad decision.
Comparisons Are Based on Unit Price, Not Total Amount
Psst's price comparisons are based on unit price, not the total line amount. For example, if you bought 2 kg this time versus 1 kg last time, the total will naturally be higher — but that doesn't mean the price went up. Only by converting to a per-kilogram unit price does the comparison accurately reflect whether prices have changed. This directly addresses one of the most common cognitive traps in everyday price comparison.
Why These Details Matter
The market isn't short on shopping list and expense-tracking apps, but most either rely on manual input or jump to hasty conclusions when auto-scanning. Psst's differentiation isn't about feature volume — it's about an obsession with data reliability: traceable row-level receipt crops, no cross-unit comparisons, and unit-price-based price tracking all work together to create a trustworthy price memory system.
For families sharing a shopping list, budget-conscious consumers, or anyone who wants to track trends in their daily spending, this "quiet but reliable" product philosophy may offer more long-term value than flashy AI-driven recommendations. The focus is on not being wrong, rather than looking smart.
Summary
Psst enters through a simple use case — a shared shopping list — and uses receipt scanning, row-level traceability, and rigorous unit-price comparison to solve the high-frequency but overlooked problem of "what did I pay for this last time?" Its restrained design philosophy stands out in today's app ecosystem, where feature bloat is the norm. The app is currently available on both iOS and Android.
Related articles

DeepSeek Open-Sources DeepInfer: What Is an Inference Framework Actually For?
DeepSeek open-sources its inference framework DeepInfer under Apache 2.0, supporting DeepSeek V4.1 Flash offline inference. This article explains its request scheduling, FlashMLA/DeepGEMM compute optimization, DSPark speculative verification acceleration, and what it means for developers and ordinary users.

Open-Source AI Learning Agent: A Full Workflow Experience from Topology Graphs to Exam Generation
Explore an open-source AI learning agent: built on DeepSeek API, it constructs knowledge systems via topology graphs, with automatic document generation, notes, super documents, and exams. Learn about its Agent workflow and design highlights.

DeepSeek Whale Girl Open-Source MV Meme: AI Industry Coded Slang About Pivoting to Closed Source
Bilibili's DeepSeek Whale Girl rap MV uses fictional memes to tease the AI open-source vs. closed-source debate, with lyrics full of industry slang. This article explains the meme culture, the Whale Girl symbol, and the open/closed tension.