mikelikesdesign

mikelikesdesign / AI-experiments

Public

Experiments on ways to utilize AI in applications

45
1
85% credibility
Found May 24, 2026 at 49 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Swift
AI Summary

This repository contains a collection of iOS app experiments by designer Michael Lee (@mikelikesdesign) exploring creative ways AI could reshape user interfaces. Each experiment demonstrates a different interaction concept - from pinching articles into summaries, to dragging sliders that change text tone, to a 3D globe visualization. The projects are self-contained SwiftUI apps that work without any external services or setup. The author shares demos publicly on social media and the code is well-documented with clear visual demonstrations. These are design explorations meant to inspire new interaction paradigms for AI-powered experiences.

How It Works

1
👀 You discover something new

You see a demo online showing someone pinching an article into a summary, or dragging a slider to change how a message sounds.

2
📱 You download and open an experiment

You pick one of the interactive apps and launch it on your phone. Everything works right away, no setup needed.

3
✋ You try the gesture

You pinch, drag, or swipe and watch the content transform before your eyes - simpler text, different tones, or a condensed summary.

4
You explore different ideas
📖
Reading experiments

Pinch articles into summaries, track your reading progress with gentle encouragement, or scroll to choose simpler or deeper content.

✍️
Writing experiments

Drag a slider to make your message sound fun or professional, or select text and rephrase it with a quadrant control.

🎮
Visual experiments

Spin a globe made of letters, or switch AI models by dragging a VHS tape into a portal.

5
💡 You see what's possible

Each interaction gives you a glimpse of how touching and manipulating content could feel natural and fun.

🎉 You understand new possibilities

You've experienced creative ways AI could become part of everyday interactions, not just chatbots.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 49 to 45 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is AI-experiments?

A collection of SwiftUI experiments exploring creative ways to interact with AI through gesture-based interfaces. Instead of traditional chat windows, these demos use pinch gestures, sliders, and spatial controls to reshape text, switch AI models, and navigate content. The standout piece is a VHS tape carousel for selecting different language models, complete with a portal animation that sends your chosen tape into deep space. All experiments run without requiring API keys, so you can explore the interaction patterns immediately on iOS.

Why is it gaining traction?

The hook is originality. Most AI UI experiments focus on prompting tricks or API wrappers, but these demos treat AI as a material to shape through physical gestures. The pinch-to-summarize feature (inspired by Arc Search) folds articles into AI-generated overviews with a satisfying accordion animation. The content rephrasing tool uses a quadrant pad that feels like mixing audio tracks instead of clicking dropdowns. These are interaction designs that make you think differently about how touch interfaces could work with generative content.

Who should use this?

iOS developers exploring novel AI interaction patterns will find the most value here. UX designers researching gesture-based interfaces for AI products can pull interaction ideas. The experiments are too experimental for production use, but they serve as a creative sandbox for thinking beyond conventional chat interfaces. If you're building anything where users manipulate AI-generated content on mobile, this repository offers unconventional starting points.

Verdict

At 45 stars and 0.85% credibility, this is a personal creative playground, not a production library. The code quality is solid with proper test coverage, but the experiments are intentionally rough prototypes. Worth exploring for inspiration if you're designing AI-native mobile interfaces, but don't expect drop-in components or documentation. The ideas are the product here.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.