Hello, I am an independent developer in Argentina building an app for rideshare drivers, and I would like to know what the supported approach is on iOS before I build the wrong thing. The problem When a trip offer appears, the driver sees a fare, a distance and a duration. What they do not see is what is left after the cost of running their own car — fuel, maintenance, and very often the weekly rent they pay for the vehicle. Here those costs are high and change constantly, and drivers routinely accept trips that lose them money without realising it. They have about three seconds to decide, while driving. Our app reads the numbers already visible on the driver's own screen, subtracts that particular driver's cost per kilometre, and shows one figure: what this trip actually leaves them per hour. In our own field measurements with two drivers over several weeks, being able to see that number raised their hourly earnings and — just as important — kept their eyes off mental arithmetic at the wheel. What we will not do, on purpose I want to be explicit about this, because I know other tools in this category cut corners: We never ask for the driver's Uber or DiDi username or password, and we never connect to or automate their account. Some tools do, and that puts the driver's account — their livelihood — at risk of deactivation. We will not build that. We never accept or decline a trip for the driver. The decision and the tap are always theirs. We never store passenger data: no name, no rating, no photo, no exact address. Text recognition runs entirely on device, offline. The captured image is discarded immediately and never leaves the phone or reaches any server. No advertising, no data sales, no third-party trackers. The app does not read anything the driver is not already looking at, it only reads while the driver has explicitly turned it on for their shift, and it exists for one purpose: so that someone working long hours knows what an hour of their work is actually worth. My question On Android this is done with MediaProjection, with explicit user consent each session and a persistent system indicator. On iOS I understand a Broadcast Upload Extension is not intended for background processing, and that there is no overlay across apps. Is there any supported path on iOS for an app to read, with the user's explicit and repeated consent, text the user is already looking at on their own screen, in order to give that same user an immediate assessment of their own work? Accessibility APIs, a system extension point, anything. If there is no supported path, I would rather know now and design the iOS version around what is allowed, instead of shipping something that gets rejected or that quietly breaks the rules. Thank you.