Stack Eye Expands from Flutter to Full Stack Detection on Any Installed App
Competitor tech-stack teardowns, 16KB compliance audits, and Flutter package adoption checks now happen in a single on-device scan, no root or separate tooling required.
Stack Eye, previously a Flutter-only detector, now lists every installed app and tags each with the frameworks and native technologies it uses. That includes 16 frameworks like React Native, Unity, Cocos, Godot, Qt, and Compose Multiplatform, plus native markers such as Jetpack Compose, NDK, Rust, and Go. The detail page adds a native-library card that checks every .so for 16KB page-size compatibility, categorizing each as ready, not ready, or not applicable, and flags missing 64-bit libraries. Users can also analyze local APK, APKS, or XAPK files from the file picker, share sheet, or file manager.
Flutter apps get an extra diagnostics card showing rendering engine (Impeller or Skia), Dart vs. engine code size, obfuscation status, and version comparison against the latest stable. App details now include SDK version, permission protection levels, exported components with intent filters and call examples, third-party SDK detection with remote-updatable rules, and screen-adaptation flags. A dashboard aggregates framework distribution, AGP versions, 16KB readiness ratio, and 64-bit support.
Reports export as images or Markdown, including a dependency-usage view that lists which installed apps use a given Flutter package, with Google, Tencent, and Alibaba highlighted for README-friendly adoption proof. All analysis stays on-device, requiring only the read-installed-apps permission. Scanning is also faster: it syncs version tables with a five-second timeout, then parses only newly installed or updated apps on subsequent launches.
Android 15's 16KB page-size requirement gives the new .so check immediate practical value; the ready/not-ready flags map directly to whether an app can ship on new hardware.
The shift from Flutter-only detection to a general tech-stack scanner reflects that modern Android apps rarely stick to one framework, so framework-agnostic reconnaissance becomes the useful baseline.
Exporting a dependency-user list as Markdown turns a local scanning tool into a promotional asset for package maintainers, letting them show real adoption numbers in their README.