Small Face Camera vs OnDevice AI Image Studio

Recognizes face from a photo and a video and automatically converts it into small face! With super easy operation, you can change to a small face and upload it to SNS and etc. This is an app that can change the face size of a photo and a video with one touch. Not only small faces, but also large faces like anime characters and loose characters can be created. Even if there are multiple faces, you can switch the face to be converted in order just by touching the screen. [Source] - Camera shooting: video, photo - Camera Roll: Video and photo loading [Configuration] - Face size: Adjust with slider - Distortion correction range: Adjust with slider - Storage resolution: 20-100% of original resolution [output] - Save to Camera Roll: Video, Photo - Video and photo sharing: AirDrop, LINE, Facebook, Dropbox, Instagram, ... [Note] If there is a geometric object on the background near the face, the background area may be distorted. In that case, adjust the distortion correction range. Ideally, photos and videos shot on a background with no geometric pattern.

Generate realistic images privately on your iPhone and iPad.  Everything runs on-device. No accounts. No uploads. Your prompts and images never leave your device. Unlimited inspirations. Why on-device * Private by design: nothing is sent to servers. * Reliable offline: generate anywhere, even in airplane mode. * Fast and responsive: optimized for Apple silicon. Create with control * Prompt + Negative prompt for clean results. * Presets and reproducible Seeds. * Adjustable Steps, Guidance, Width, Height. * One-tap Share to Photos or any app. Designed for clarity * Single-screen layout on iPhone and iPad. * Colorful, accessible UI with large controls. * Works in light or dark appearance. What you can make * Photorealistic people and places. * Product mockups and concept art. * Backgrounds, wallpapers, and social images. Privacy * No account, tracking, or analytics. * No data collected. All generation happens locally. Notes * First run may take longer while the model loads. * Results depend on prompt quality and device performance.