OP here. A bit more context: I built this mainly for packaging cross-platform release artifacts where ISO compatibility gets tricky (Unicode filenames, files >4GB, and preserving UNIX permissions).
The app can create new images and also modify existing ISOs (add/replace/delete files, then rebuild). It supports ISO9660/Joliet/UDF/Rock Ridge so you can pick a compatibility set that matches your audience.
I’d love feedback on real-world edge cases to test: mounting behavior across OSes, Unicode normalization, deep paths/long filenames, large directory trees, etc.
two quick notes that might help you evaluate it faster.
1 heatmaps and reader paths are built from session-level page transitions, not just aggregate views.
2 the viewer does not need an account, and recipients do not need to log in.
i would love blunt feedback on two things.
which signal would change your next follow-up step. drop-off points, re-reads, entry pages, or something else
what feels missing compared with how you personally review proposals or decks
happy to answer technical questions about the tracking model and how we deal with noisy sessions.
i get the appeal of vibe coding, but in practice it often shifts the hard work instead of removing it. you write less code, but you spend more cycles reconciling what the AI produced with how the system actually works. for small experiments it’s great, but once a product has real users, the cognitive load comes back with interest. the speed is real, but so is the debt.