The week of June 6–12 at Rork Lab centered on a single idea: the decisions that only start to matter after you ship. In the previous weekly highlights I mentioned wanting to shift toward post-launch operations around billing, and this was the week that plan finally took shape. Subscription refunds, wallpaper-app binary size, deep links that silently fail on cold start — none of these show up on a pre-launch checklist, yet every production app eventually has to face them.
Alongside the hands-on pieces, I also published two articles written from a step back: a candid review of six months running Rork, and a break-even guide to Rork Max pricing. As always, everything here started from something I actually ran on my own devices.
Design as if refunds never reach your app — because they don't
The article I'm most glad to have written this week is the premium piece Your App Never Notices Subscription Refunds — Revoking Entitlements with REFUND Notifications and Voided Purchases.
When a refund goes through, the Premium flag inside your app stays set. The user keeps their money and keeps the paid features — a structural gap rather than a bug. I'll admit I only noticed it in my own apps after the billing flow had been live for a while: I had never explicitly decided how entitlements should be revoked. The article organizes the revocation paths into three architectures — App Store Server Notifications V2 with its REFUND type, Google Play's SUBSCRIPTION_REVOKED plus the Voided Purchases API, and RevenueCat's listener with a launch-time reconciliation — and finishes with deny-by-default on the client as the last line of defense. There's also a section on reducing refunds in the first place. If your app sells subscriptions, it's worth checking whether a revocation path exists at all.
Twelve years of wallpaper apps, and redrawing the binary-size boundary
The second premium article is Keeping Wallpaper App Binary Size Under Control — When to Move Images Out of the Bundle.
Wallpaper apps grow heavier with every batch of content by their very nature. Running Beautiful HD Wallpapers over the years, I've redrawn the line between bundled and remotely delivered images more times than I can count, and reworked the prefetch design just as often to keep first paint fast. This article consolidates those decisions with real measurements, and answers the "should I use On-Demand Resources or Play Asset Delivery?" question from the angle that matters for solo developers: ongoing operational cost. One finding from measuring it all properly — trimming size ripples into more than review times and install conversion.
Deep links and Expo offline starts — leaving behind the diagnosis, not just the fix
Implementing Deep Links in Rork Max — Universal Links and URL Schemes walks through carrying a user straight to the right screen from a single tap: when to use URL Schemes versus Universal Links, the AASA and assetlinks pitfalls, React Navigation integration, and the cold-start cases where links silently vanish. Deep links are a domain where "configured but not working" is the default failure mode, so the article is explicit about what to suspect first.
expo start --offline Fails with forbidden — Separating Corporate Proxy 403s from Dependency Validation deals with one error message that has two unrelated causes: an HTTP 403 returned by a corporate proxy, and the Expo CLI's validateDependenciesVersions guard. I included three dependency-version conflicts I actually hit in my own projects. It's an unglamorous topic, but it's the kind of article that can save a few hours for whoever lands on it from a search.
Two pieces from a step back: the six-month review and the pricing math
What Has Rork Actually Changed as an 'AI Mobile App Builder'? — Evaluation Criteria from Six Months in Production opens with why the word "builder" never quite fit, then maps where Rork genuinely helps, where it stalls, and how it really differs from Lovable, Bolt, and FlutterFlow. It closes with three questions that decide whether Rork is the right starting point, plus a June update on Rork Max's native Swift generation.
Rork Max Pricing Decision Guide 2026 — Choosing a Plan by Project Type and Break-Even Point breaks the pricing into fixed monthly cost plus execution-based billing, and runs the break-even math for three profiles: weekend side-project developers, indie developers with recurring revenue, and agencies. Pricing articles tend to collapse into feature comparisons, but having run multiple apps solo, I find the earlier question is whether the billing model fits your development rhythm at all.
Looking ahead
Writing the refund article reinforced something I keep relearning: billing quality is determined not on the day you implement it, but on the day something unexpected happens. Next week I plan to dig into review-response and localization operations, along with techniques for maintaining Rork Max-generated code over the long term.
Thank you for spending part of your week with Rork Lab. I hope your apps run quietly and steadily in the week ahead.