The Opportunity
Emotional clarity matters more than ever. One in six people worldwide experiences loneliness, while one in ten U.S. adults retrospectively reported childhood emotional neglect. Research involving 120 adults also found that five days of emotion education improved participants’ ability to distinguish between negative emotions. Employers now rank creative thinking, self-awareness, and empathy among the most valuable workplace skills.
How Bearings Works
Bearings is designed for three brief check-ins each day: one in the morning, one around midday, and one in the evening. Users write about how they feel in the moment, then use on-device AI to explore their inner world more precisely. When a user taps the magnifying glass, the AI identifies vague language and asks thoughtful questions that encourage deeper reflection.
As entries accumulate, Bearings creates a private summary of each day and week. At the end of the week, the app provides a reflection and a practical lesson to support continued emotional growth.
The Design
I used ChatGPT throughout the design and development process to help turn the idea into a native iPhone app. Bearings uses Apple’s on-device Foundation Models framework to generate reflections without sending journal content off the device.
- Swift and SwiftUI power the interface.
- SwiftData stores journal entries and reflections locally.
- Foundation and Calendar preserve dates across time-zone changes.
- Swift concurrency manages saving, cancellation, and AI tasks.
- Apple Foundation Models provides on-device language generation.
Valleys and Ravines
Early versions of the concept felt limited to recording journal entries and presenting general psychology lessons. The central challenge was finding a way to make reflection more personal and useful. On-device AI helped solve that problem by identifying vague language, asking targeted questions, and focusing each reflection on the emotional skills the user may need to develop.
Milestones & Achievements
The on-device AI is the part of Bearings I am most proud of. Building a locally running model experience that feels genuinely useful required extensive experimentation and testing. I also handcrafted the interface to feel calm and tactile, drawing inspiration from topographic maps, old kerosene lamps, parchment, and fountain pens.
Building Bearings showed me how much potential Apple’s Foundation Models have beyond this first version of the app. This was also my first time using Figma, collaborating with ChatGPT across design and development, and competing in a hackathon. I am grateful for the experience and excited to hear what people think of the project.
The Future Development of Bearings
Next steps include launching a TestFlight beta, simplifying the interface further, and refining the AI through better prompts and evaluation. I also hope to expand its educational approach using ideas from books that have shaped my own journey, including Voice of the Heart, Running on Empty, The Anxious Generation, and The Language of Emotions.
If you are interested in testing Bearings, contact me at [email protected].
Log in or sign up for Devpost to join the conversation.