Selah

Problem

My girlfriend was going through a difficult time. CBT teaches that emotions are shaped not by events directly, but by the thoughts that interpret those events. I wanted to build something that could help her — and eventually others — identify cognitive distortions and reframe their thinking.

My Role

I designed, built, and grew the product entirely solo — product strategy, AI integration, UX design, React Native development, user research, and community engagement.

Process

• Built an initial CBT tool using the Claude API to guide users through identifying cognitive distortions and reframing negative thoughts. Used it together with my girlfriend as the first user.
• A friend tried the app and said he'd actively use it as a product. Since he's a practicing Christian, I hypothesized that focusing on a specific audience — people seeking both emotional and spiritual support — would be stronger than building a broad mental health tool.
• Integrated personalized Bible verse recommendations and redesigned the experience for users at the intersection of mental health and faith.
• Engaged directly with church communities — observing how people share prayer requests and support each other weekly. Translated this into an anonymous 'prayer room' feature where users could share and support each other online.
• Grew to ~100 weekly active users through community engagement and ran a small church pilot to validate the product in real-world settings.
• Built the full stack: React Native + Expo frontend, Supabase/PostgreSQL backend, Claude API integration for AI-powered thought analysis.

Outcome

~100 weekly active users with strong engagement and retention. Users expressed willingness to continue using the product if launched. Qualitative feedback highlighted the trust and authenticity of the experience. Now preparing for launch with monetization.

What I Learned

The hardest part was designing a product that handles sensitive emotional experiences while maintaining trust and authenticity. You can't fake empathy in UX — every word, every screen transition, every notification matters when someone is vulnerable. I also learned that the best product decisions came from going narrow: targeting a specific community instead of building for everyone made the product stronger, not weaker.