A mobile bookmarking app with AI shopping recommendations

Shoplater is an internal Twistag product — a cross-platform mobile bookmarking app for shoppers who save now and decide later, with an AI layer that surfaces saved items when they are worth acting on. It is Twistag's own product, not client work. Shoplater solves a category problem: browser bookmarks do not survive across devices, wishlists are siloed per retailer, and screenshots get lost in a camera roll. Twistag built it on React Native for iOS and Android from one codebase, with a Python parsing layer on AWS that extracts product details from any retailer URL without needing that retailer's cooperation, and item-level conflict resolution for cross-device sync.

Shopping decisions do not happen in a single session. A shopper sees a product they like at lunch, gets distracted, and never returns to it. Browser bookmarks do not survive across devices. Wishlist features are siloed per retailer, so every store keeps its own list and nothing is comparable across them. Screenshots get lost in a camera roll. The thing the shopper wanted to remember disappears into the day. Shoplater is an internal Twistag product, built and owned by Twistag as a consumer-facing testbed for its product engineering and AI recommendation work. It is not a client engagement and is not counted in Twistag's client project totals. The build had two hard parts — a fast, reliable cross-platform mobile experience, and a recommendation layer good enough that returning to saved items feels like discovery rather than housekeeping.

Twistag built Shoplater on React Native, shipping iOS and Android from a single codebase without giving up native feel. The save flow lives in a share extension on both platforms, so a shopper browsing any retailer's app or website adds a product in one tap — a deliberate decision to work universally rather than to depend on per-retailer partnerships, because negotiating API access with thousands of retailers is not a viable path for a consumer product.

Behind that, Twistag built a parsing layer in Python on AWS that extracts product details — name, price, image, retailer, availability — from a wide range of source URLs and page formats, with Open Graph metadata as the fallback path when a retailer changes its layout.

The recommendation engine is where the product earns its name: Twistag designed it to surface saved items contextually rather than chronologically — items that have dropped in price, items matching behaviour patterns such as more interest in casual wear before the weekend, items saved repeatedly across categories. Cross-device sync was treated as a core product feature rather than an afterthought, built on AWS with conflict resolution at the item level so actions stay coherent across devices even with intermittent connectivity. Working universally rather than through partnerships is the decision that defines the product. A bookmarking app that only saves from retailers who agreed to an integration is a wishlist with extra steps, so the parsing layer had to hold against layouts Twistag does not control and cannot be notified about when they change.

Stack: React Native, Python, AWS. Services: Product engineering, mobile engineering, AI recommendation systems, product design.

Shoplater works as the consumer-facing testbed for Twistag's product engineering and AI recommendation capabilities, and demonstrates three patterns Twistag applies to client work: cross-platform mobile delivery from a single codebase, AI recommendations that read as useful rather than intrusive, and treating sync and offline behaviour as core product features. Lessons from the recommendation engine and the cross-device sync architecture have flowed directly into Twistag client projects in retail and DTC. No usage, revenue or retention metrics are published for Shoplater, and this page does not claim any.

The recommendations don't just remind users what they saved — they suggest when to act on it.