AI Application
Screenshot AI App
An AI-powered iOS concept that turns screenshots into structured, actionable tasks.
Overview
Screenshot AI App is a concept for an iOS application that turns any screenshot into a structured, actionable task — reading the image with AI and extracting what matters, so a screenshot becomes something you can act on instead of just another photo in your library.
The Idea
Screenshots are one of the most common ways people capture information — a flight confirmation, an address, a task someone sent in chat — but that information usually stays trapped as a flat image. The idea is to close that gap: point the app at a screenshot and get a clear, structured task back.
My Role
Defined the product concept and UX flow — how a screenshot becomes a structured task — designed the AI interaction and prompt design, and planned the iOS app structure and integration. AI-assisted development is used throughout the process.
AI Usage
OpenAI's vision and language models are the core of the concept, explored for reading a screenshot and turning it into a structured, actionable task. Claude is used for prompt iteration and AI-assisted development of the prototype.
Development Process
- 01Concept: define what a screenshot-to-task pipeline should feel like
- 02Prompt design: iterate on how AI reads and structures a screenshot
- 03Structure: plan the iOS app flow around the pipeline
- 04Prototype: begin building toward a first working version of the pipeline
Challenges
- Getting structured, reliable output from an AI model reading unpredictable screenshot content
- Designing a UX flow simple enough to feel instant, not like configuring a tool
What I Learned
- Prompt design has as much impact on reliability as the underlying model
- Structuring AI output for a UI is a distinct design problem from getting a good one-off answer
- Small iOS concept apps are a fast way to test whether an AI interaction actually feels useful
Outcome / Current State
Prototype stage: the core screenshot-to-task pipeline has been explored and works as a concept. The iOS app shell and broader task management around it are still open. Next step is a working end-to-end prototype on-device.