I built Zup, a free native iPhone app that turns photos and ideas into social posts. The part I care most about is what happens after the AI produces a draft: you should still recognize your own voice and control what gets published.
Keep the draft editable
Zup lets you start with photos or a text-only idea, choose preferences for tone, length, emojis and hashtags, then edit the caption in a live platform preview. Proofreading is available without asking you to start the whole draft again.
That means generated text is a starting point in the writing workflow. The person sharing the post still reviews the wording and chooses what to keep.
Keep photo edits reversible
The native SwiftUI app includes photo cropping and editing while preserving the original. I wanted the same idea to apply across writing and images: trying an edit should not mean losing the version you liked before it.
Make scheduling an explicit decision
Recent updates added scheduling for X and LinkedIn. You review a post and confirm its publishing time; the scheduled publish runs on the server, so the app does not need to stay open. Other sharing paths depend on platform and account support, with the iPhone share sheet available as a fallback.
One useful distinction when building this kind of mobile flow is that preparing content and publishing it are different actions. A good preview helps people catch the wrong wording, destination or image before they commit.
Be clear about the AI boundary
Zup asks permission before hosted AI processing. It is not an entirely on-device AI app. Keeping that distinction clear matters when people are choosing photos to send for processing.
All features are free, with an optional one-time tip.
You can explore it at zupapps.com or download it for iPhone.
If you build mobile writing tools, where would you put the final review step: inside the editable preview, or on a separate confirmation screen? I would also appreciate feedback on any part of the photo-to-post flow that feels unnecessary.
Top comments (0)