How CartLens Turns In-Store Receipts into Price Transparency for Offline Shopping

I felt this every single time I went shopping. I'd walk into a store, pick something up, pay for it, and have no real way of knowing if I'd actually gotten a good deal or just gotten taken advantage of. More than once, I'd shop at one store, then walk into another one nearby a little later and find the exact same item for less, after I'd already paid, when there was nothing I could do about it anymore. That timing always bothered me the most. Online, you can compare prices across five sites in the time it takes to open a browser tab. Walk into a physical store though, and all of that just disappears. You're stuck trusting the sticker and hoping it's fair. I wanted something that could close that gap. Something that could tell me, right there in the aisle, before I paid, whether what I was about to spend actually made sense, instead of finding out the hard way afterward.

Right now, what's actually working is getting listed on relevant directories and sites in our niche. It's a pretty simple move, but it puts us in front of people who'd genuinely never have found us otherwise, since they're already out there looking for something like this. We're also posting on social media pretty consistently, day to day, though honestly, it's still too early to say with any confidence whether that's turning into real usage or just visibility. Local influencer partnerships are next on the list. I think that could really move the needle, since a trusted local voice carries a lot more weight than an ad ever will, but that takes a marketing budget I don't have available right now, so it's on hold until funding comes through. Honestly, the project is young enough that I can't point to anything that's clearly flopped yet. Right now, the real work is less about avoiding failures and more about figuring out which channels are actually worth doubling down on as we get more data.

The biggest complaint early on was friction, plain and simple. The only way to actually track what you were paying over time was to manually type every item, size, and price into a spreadsheet, and keep doing that after every single shopping trip. Nobody has the patience for that, especially not after standing in line and unloading a cart. It's the kind of thing that sounds useful in theory but that everyone quietly abandons after a week or two because life just gets in the way. So instead of trying to make the spreadsheet less painful, we solved it at the source. You just point your phone at a receipt or a shelf tag, and the AI reads out the item, the size, and the price automatically, in seconds. No typing, no spreadsheets, no remembering to log anything later. Just a photo, and the data's already there waiting for you. That was really the whole point, remove the effort instead of just relocating it.

Everyone assumes this is a grocery problem. It's not; grocery's just the easiest example to point to because everyone shops for food. The truth is the same blind spot shows up anywhere you buy something in person, pharmacy, hardware, electronics, clothing, all of it. The bigger misconception, though, is around delivery apps. People assume that because an app shows them a price, that price reflects reality, the actual cost of the item on the shelf. It doesn't. A lot of delivery platforms mark up the listed price anywhere from 15 to 30 percent before delivery fees, service fees, or tips even get added on top. So, what looks like a fair, transparent price on your screen is often quietly inflated before you've even started checking out. If you actually want to know whether you're getting a good deal, the app isn't where you find that answer. You have to look at what's really on the shelf, in the store, where the price actually starts.

We're starting really local, in the SF Bay Area and Oakland, aiming for 20+ stores per zone before we expand anywhere else. That's how we actually solve the problem of starting from zero data, instead of claiming thin coverage everywhere and being accurate nowhere. We're partnering with local budgeting creators on same basket, different store price audits, and seeding early receipt contributors through neighborhood and parenting groups. We've also set aside a $15K pilot budget to validate signup costs, receipt submission rates, and free to Pro conversion before we scale further. Once that data proves out in one metro, the plan is to repeat the same playbook city by city rather than rushing into a national push.

Over 100 people have signed up so far, but zero of them are paying customers yet.

I'm not new to building things. I spent years working as a full stack developer and software architect, and eventually ran my own agency, building everything from web applications to AI integrations and blockchain projects for a range of different clients. So the technical side was familiar ground. What was genuinely new was applying that experience to physical retail specifically. Roughly 68% of retail still happens offline, in person, at an actual register, and yet almost none of the data infrastructure, automation, or AI-powered transparency I had spent years building for online businesses existed for that offline world at all. That gap, between what I already knew how to build and what simply didn't exist yet for physical shopping, is basically where CartLens came from.

My biggest mistake was building before validating anything. I went straight into development without properly studying the market first, without clearly figuring out who my real competition actually was, and without stopping to ask whether the problem I saw was one other people felt just as strongly. I ended up spending far more time writing code than talking to potential customers, which is backwards. What I learned is that customer conversations are what actually tell you what to build. Skipping them doesn't save time; it just means you're guessing, with extra steps and extra code added on top of the guess. Now, before I write a single line of code for a new idea, I make sure I've talked to real people about it first.