Inspiration
PLens came from a simple problem I experienced after moving to Portugal: shopping becomes surprisingly difficult when you don’t understand the language on the products.
Something as simple as buying milk can turn into a guessing game. Is it whole or semi-skimmed? Is it lactose-free? What percentage of fat does it contain? The same problem appears everywhere — toothpaste, mouthwash, detergent, shampoo, food, and countless everyday products.
Traditional translation apps can translate the words, but that’s not always what you need. When standing in a store, you don’t want to translate an entire package and then figure out which information actually matters. You just want to know:
What is this, and is it the one I’m looking for?
That’s what inspired PLens.
What it does
PLens is a camera-first product interpreter designed for travelers, digital nomads, expats, international students, and newcomers living or traveling in countries where they don’t understand the local language.
Point your camera at a product and PLens identifies it, translates it, and extracts the information that actually matters.
For example, instead of simply translating a Portuguese milk carton word by word, PLens can tell you:
Lactose-Free Semi-Skimmed Milk
- Lactose-free
- 1.5% fat
- Cow’s milk
- UHT
- 1 L
What PLens shows changes depending on the product. For mouthwash, that might be whether it contains alcohol or fluoride. For detergent, it might be whether it’s intended for colors or whites, its recommended temperature, and the number of washes.
The goal isn’t just to translate packaging. It’s to understand it.
The core experience is:
Point → Understand → Choose
Users can also save products they discover, making it easy to remember and find products they liked again.
How we built it
PLens is being built as a native iOS app with a strong focus on speed, simplicity, and a polished camera-first experience.
The product-analysis pipeline combines native iOS camera and vision capabilities with multimodal AI. Rather than asking AI to simply translate visible text, PLens uses structured product understanding to determine what kind of product the user is looking at and which attributes are most useful for that particular category.
The AI response is converted into structured information that the interface can present quickly and consistently instead of showing the user a long AI-generated paragraph.
The architecture is also designed to minimize unnecessary AI processing through efficient image handling, local processing where appropriate, and caching previously analyzed products.
RevenueCat powers PLens’s premium access and monetization.
Challenges we ran into
The biggest challenge with PLens isn’t translation — it’s deciding what matters.
A milk carton, toothpaste box, shampoo bottle, and laundry detergent all contain completely different information. PLens needs to first understand the context of the product and then determine which information is actually useful to someone deciding whether to buy it.
Accuracy is another major challenge. Information such as whether something is lactose-free, alcohol-free, or suitable for a particular use shouldn’t be confidently presented based on an unsupported AI assumption. The experience needs to distinguish between information supported by the packaging and uncertain interpretations.
Speed is equally important. PLens should feel like pointing a camera at something, not submitting a question to an AI chatbot and waiting for an answer. Creating that feeling while using multimodal AI is one of the most interesting technical challenges of the project.
Accomplishments that we're proud of
We’re proud of keeping PLens focused on one clear real-world problem instead of turning it into another general-purpose AI scanner.
The experience is intentionally simple: point at something unfamiliar and understand it.
We’re also proud of making the information category-aware. Instead of returning the same generic analysis for every scan, PLens focuses on the attributes that matter for the specific product being viewed.
Most importantly, PLens turns a capability that can feel technically complex — multimodal AI, image understanding, translation, and structured extraction — into an interaction that requires almost no explanation from the user.
What we learned
Building PLens reinforced that translating text and understanding a product are two very different problems.
The most useful answer isn’t necessarily a complete translation of everything written on the packaging. Often, users only need a few pieces of information to confidently make a purchasing decision.
We also learned how important speed, structured AI output, and careful handling of uncertainty are when bringing multimodal AI into a real-world camera experience.
The best AI experience isn’t necessarily the one that shows the user the most information. It’s the one that understands which information matters right now.
What's next for PLens: Smart Product Scanner
Our vision is to make PLens the fastest and easiest way to understand any product, anywhere in the world.
Next, we want to make product recognition faster, smarter, and more flexible. Beyond pointing the camera at a product, PLens will support additional ways to identify products, including barcode scanning and importing product photos directly from the photo library.
We’re also planning a Deep Dive mode for moments when users want more than the essentials. While the default experience will remain fast and simple, Deep Dive will provide richer information such as ingredients, nutrition, allergens, usage instructions, warnings, and detailed translations when relevant.
Saved products will become more useful too. Users will be able to organize products they discover, add personal notes, revisit products later, and share products with friends and family.
Further ahead, we want PLens to help users compare similar products, find local alternatives to products they already know and love, and naturally learn useful words from the products they encounter while living or traveling abroad.
The core experience, however, will stay simple:
Point. Understand. Choose with confidence.
Built With
- avfoundation
- ios
- ocr
- rest
- revenuecat
- storekit
- swift
- swiftui
- vision
Log in or sign up for Devpost to join the conversation.