mikelikesdesign/ai-experiments

This repository is a personal collection of interaction design experiments that explore different ways to weave AI into app interfaces. Each experiment is a small, focused prototype that tries out a specific gesture or interaction pattern, such as pinching, dragging, or tapping, as a way to control AI-powered content changes. Many of the experiments focus on reshaping text on the fly. A pinch gesture compresses an article into a summary. A drag slider rewrites a message across different tones: concise, detailed, professional, or fun. Another drag control shifts reading content between a simplified version and a more in-depth one. A word-tap experiment lets the user press and hold any word in a passage so the AI surfaces synonyms, then tap one to swap it in. Another prototype watches the user's facial expression while they read, if they look confused, the AI shows a simpler version, and if they seem interested, it offers more detail. Other experiments explore navigation and discovery. One creates a visual diagram of connected topics based on a word or phrase the user types. Another adds a prompt navigator to a chat history, revealed by a pinch gesture, so the user can jump back to earlier exchanges. A reading tracker responds with small encouragement messages as the user works through an article. The author notes that most of the demos have been updated to run without an AI API key, so you can install and explore the ideas directly. If you want a version that connects to a live language model, the README directs you to contact the author on X. Several experiments listed in the README have their code marked as coming soon, meaning the source has not been posted yet.