Designing a wellness experience around emotional safety, autonomy, and reflection rather than performance.
A mobile app for people who find tracking stressful. Researched, tested with 15 people, and built as a working React Native app.
Wellness apps often turn wellbeing into a score.
Tracking can build awareness. But for people dealing with anxiety, depression, eating disorders or low motivation, the same mechanics can create pressure, guilt and a sense of failure.
The tools meant to promote wellness can sometimes undermine the mental health they are trying to support.
The design questionHow might we design a wellness experience that encourages healthier habits without turning wellbeing into another performance metric?
Secondary research: a literature review on self-tracking and eating-disorder risk, an analysis of how popular apps engage their users, and behavioral psychology and ethical-UX guidance. It produced four insights that shaped everything after.
Logging can make people more mindful, but an emphasis on calorie counting and weight tracking can create pressure, anxiety and guilt. Use of weight and fitness management apps has been correlated with disordered eating behaviors.
Eikey, 2021 · Huang et al., 2022
Streaks and badges can encourage compulsive behavior, discourage rest days, and push people to carry on through illness to keep a number going.
Feature analysis · Landsverk, 2023 · Cherry, 2020
Apps that focus on emotional support and understanding tend to support lasting improvements. Usability alone isn’t enough when the user may be struggling.
Eikey, 2021
Self-Determination Theory points to autonomy, competence and relatedness as basic needs. For apps, that means letting people make their own choices instead of following strict rules.
Deci & Ryan, 1985
Three principles turned the research into constraints I could design against.
Design for how users feel, not just what they accomplish.
A soft, rounded interface with warm, non-directive copy. Crisis resources are one tap away (“Get Immediate Help”). The AI tool carries explicit boundaries.
Users should control how and when they engage with the product.
Users choose whether to see calorie and macro info, which widgets to pin, how the app supports them from onboarding, and whether to use the AI tool at all.
Encourage awareness and sustainable self-care rather than optimization.
Meal reflections capture fullness and mood. Insights show mood, sleep and energy in gentle language. Optional prompts help people notice non-numerical progress.
This is the working app, recorded on a phone. Pick a chapter to jump to it.
I didn’t just design this. I built a working product.
Calorie counts can lead to fixation and guilt, especially for people with eating disorders.
Track how food feels, not only what it is: qualitative tracking such as hunger cues and emotional check-ins. Numbers become a choice.
Log Meal asks for fullness (1–5) and mood after eating. Preferences ask whether to see calorie or macro info at all.
Rigid metrics and streaks can make people feel they are failing when they can’t keep them up.
Flexible reflection and user-controlled engagement. Progress shown in gentle language, not as a score to defend.
Insights: a mood calendar with mood, sleep and energy trends, and messages like “Your sleep has been a bit off — maybe time for extra rest?”
People have different comfort levels with support, and control is part of wellbeing.
Customization and choice: users shape their own Home, set their own goals, and choose how the app supports them.
Pinned widgets on Home, and a widget library where every tool can be pinned or left alone.
Traditional thought records can feel clinical, rigid or overwhelming.
Break them into smaller, optional prompts, in warmer, non-directive language.
Guided steps for identifying a trigger, describing a feeling, and gently exploring a more balanced thought. See the next section.
Community features can offer support, but also promote comparison and spread triggering content.
Peer encouragement over performance metrics. Light moderation, content filters, reporting tools and clear community guidelines.
A Community space in the main navigation.
How do you turn tools from clinical, paper-based settings into a self-guided experience that doesn’t feel clinical or overwhelming?
I reviewed a broad range of CBT-based tools and chose three. Each is backed by clinical evidence, and each favors self-awareness over performance or correction.
Teach users to challenge distorted thinking and reframe negative self-talk. In Vira the process is bite-sized: identify a trigger, describe a feeling, gently explore a more balanced thought.
A personal resource to return to in a hard moment. Designed as a customizable space for notes, affirmations, images or calming ideas, to give it a personal, comforting tone.
A lighter-touch tool for tracking internal dialogue. It builds awareness without judgment.
In the app these appear as Thought Reframe, CBT Thought Record, Reflection Journal and Coping Box, each a widget the user can pin to Home or leave alone.
Reflection over correction: inviting people to slow down and notice patterns, without the pressure to “fix”.
“This is the first time I’ve used a wellness app and didn’t feel worse about myself afterward.”Participant
Users weren’t sure what progress meant without traditional metrics.
I added optional reflection prompts that help users recognize non-numerical signs of growth, like better sleep, less anxiety or more emotional awareness, and define success on their own terms.
Users wanted onboarding to feel more personal.
I redesigned onboarding so users customize how the app supports them: gentle mood tracking, journaling prompts, or reminder-based encouragement.
The AI reframing tool felt like support, which risked being mistaken for therapy.
I added disclaimers, a choice of self-guided or AI-supported tools, validation-first responses and a resource list.
How do you offer supportive AI without making people believe they’re receiving therapy?
The thought-reframing tool had two modes. Testers called the conversational one “surprisingly comforting” and “like texting someone who listens”. That was exactly the problem: I had to weigh the emotional benefit of real-time support against the risk of users mistaking the AI for therapeutic care.
Complete a guided thought record independently.
Reframe a thought by talking with an agent that mimics the tone of a supportive mental health coach.
The goal was a wellness companion, not a therapist: support that empowers people without creating dependency or false expectations.
I design what I build, and I build what I design.
Wireframes and UI/UX exploration, aiming for a soft, calming interface that doesn’t overwhelm.
The working prototype, cross-platform across iOS and Android.
Secure authentication and private data storage.
Home: a greeting, a collapsible check-in and widgets the user pins.
Log a meal by how it felt: fullness and mood.
Calorie and macro info is a question, not a default.
Meal ideas framed around mood.
Movement without pressure: “every step counts”.
Coping & Comfort, with crisis help always reachable.
A mood calendar and trends, in gentle language.
Hydration with a mascot, and no scorecard.
IMA Capstone Week, Spring 2025. My original presentation of the project.
Watch on Vimeo →Empathy had to be built into every stage, not added at the end. It shaped copy, customization and safety features alongside the interface.
Choosing three CBT tools meant choosing reflection over correction, and every feature had to answer to that.
Removing metrics felt right, but it left some new users without a sense of direction. The answer wasn’t to bring the numbers back. It was optional prompts and letting people define success themselves.
The most valuable AI decisions were boundaries: disclaimers, choice, validation over advice, and a path to professional help.
A week of real use with 15 people surfaced problems a design review wouldn’t have, especially around onboarding and progress.
Shipping a working app meant the design was tested as a product, with real data, real states and real consequences for privacy.
Vira is a wellness and self-reflection tool. It is not a medical device, does not diagnose or treat any condition, and is not a substitute for professional care.