mdrng.
Habit trackers assume motivation is the problem.
Many productivity apps use up a lot of real estate convincing you to do things you’ve already decided to do. The person I was designing for isn’t looking for streaks or badges. They need a clean surface, fast entry, and honest progress tracking with no guilt attached – a simple tool that works.However, a focus on utility doesn't have to be joyless. Tools that treat simplicity as a feature often strip out any reason to return. MDRNG had to earn engagement through craft, not mechanics.
Design principles
Five principles acted as active filters throughout — each with a "this means we don't…" consequence that made them useful when things got ambiguous.Progress as a visual language
Progress is represented as a continuous gradient from deep purple to warm coral. It makes partial completion feel real and visible, and turns the dashboard into an ambient signal about your day — a page full of coral tells you something without any numbers. Don’t like the colors? I provided the user with the ability to choose a gradient that suits them.Holding the Line
Every proposed addition was tested against the principles: does this serve the intention, or does it serve engagement? The clearest example was removing an edit screen entirely in favour of inline drag-to-adjust progress. I was attached to the screen from early in the process but the value did not justify the friction it created.Earning re-engagement through atmosphere
A tool with no reason to return isn't a product, it's a form. A physics-based dashboard visualization reflects overall progress as an ambient visual animation — not a reward, not a score, just the pleasure of the thing looking different each day.AI as the connective tissue of a solo process
This project started as a fun idea without clear purpose. What surprised me was that working through it with AI helped me understand what I was building. The design principles above weren't invented; they were excavated. Some were rationale I already held but hadn't named. Others I hadn't consciously formed until the conversation drew them out.In terms of execution, AI was able to give shape to what could have been a meandering solo project. Conversations with Claude helped with everything from defining clear goalposts and next steps to identifying which screens to provide for development with Cursor.
AI acted as a fluid support structure, removing friction at every stage and providing analysis and structure whenever the need arose. It wasn't without its limits; there were moments where it pushed in directions I wasn't interested in, and maintaining authorship meant staying deliberate about the decisions that defined the product. The visual language, the product philosophy, the calls about what to cut stayed mine throughout.