Here is a technical-first Product Hunt maker comment you can post: We built FitWallpaper from a very specific pain: on iPhone lock screens, the subject I care about (idol, anime character, portrait) often gets covered by the clock. iOS depth effects and wallpaper duplication exist, but they frequently introduce blur or visible quality loss. Our approach is device-aware wallpaper generation, not generic image expansion. FitWallpaper pre-allocates composition safe zones based on real layout constraints: clock area, bottom app region, and side margins. Then we run a masking-based extension pipeline to preserve the original aspect and art style as much as possible, upscale for higher output quality, and export directly to iPhone portrait and iPad landscape (beta). A key product decision changed during development. We first tried a single-pass “one-shot perfect result” strategy, then considered hard branching by image type. Because photo distributions are highly diverse and non-linear, we moved to parameter-tuned multi-output generation so users can pick the best result per image. Current status: iPhone portrait: live iPad landscape: beta In progress: better iPad fit accuracy, faster generation, automatic subject detection, lock-screen golden-ratio subject placement research, and better landscape-to-portrait conversion (currently the hardest problem) What we want feedback on: How to push generation latency below ~1 minute without sacrificing quality Better methods for high-resolution and landscape photos to keep subjects centered without distortion If you test it, sharing your hardest photo case would be incredibly helpful for improving the pipeline.

FitWallpaper

AI-expanded wallpapers for iPhone and iPad lock screens.

AI-expanded wallpapers for iPhone and iPad lock screens.