The Mindloom

• 

  Discover a personalized mental wellness companion that helps you understand, track, and improve your emotional wellbeing.

What inspired me

Mental health is hard to talk about—especially when you're afraid of being judged. The idea that someone can open up to an AI companion anytime, without fear, changes everything. Traditional therapy often requires scheduled sessions and availability, but what if you're overwhelmed at 2 AM? That inspired me to build The Mindloom—a 24/7 mental wellness platform where anyone can check in, reflect, or talk to an AI companion without pressure. It’s private, always there, and designed for moments when you need just enough support to feel seen and steady.

What I learned

Creating emotional safety with AI is less about tech and more about empathy. I learned how important it is to design for the in-between moments—the 3-minute check-ins, the need to quickly vent, the desire to track feelings over time. It taught me that user journeys in mental wellness are deeply personal and that even small nudges (like a journal prompt or voice tone) can make a big difference.

How I built it

The Mindloom was built using Next.js 15, Supabase for the backend, and Clerk for authentication. LemonSqueezy powers the subscription layer. I integrated OpenAI for AI conversations and ElevenLabs for conversational AI. The app includes daily mood check-ins, contextual journaling, and an AI Companion that can talk, listen, and reflect.

Every feature is designed to feel fluid—whether it's a user tracking their mood for the week or jumping into a spontaneous voice session. The content is served through a markdown-powered blog system, and all user interactions are stored securely with real-time feedback loops and memory retention for a more personalized experience.

Challenges I faced

Prompt engineering was one of the biggest hurdles—how do you make an AI sound warm, helpful, and never intrusive? Managing voice sessions with ElevenLabs and ensuring smooth transitions between text and voice modes was another technical challenge. I also had to work carefully with webhooks for credit deduction and data integrity to avoid race conditions or billing issues during live interactions.

Accomplishments that I'm proud of

• Built an emotionally sensitive product that feels natural and trustworthy.
• Created a full stack voice-powered therapy flow that works in real-time.
• Designed a modular journaling and check-in experience grounded in real mental health practices.
• Developed a unique credit-based subscription system that enables flexible session usage.
• Maintained a clean, calm design system that supports focus and healing.

What's next for The Mindloom

I’m planning to bring in an AI video companion powered by Tavus.io to create deeply personal, expressive responses that feel even more human—perfect for moments when visual empathy matters.

Micro-interventions are also getting a major upgrade: using ElevenLabs' voice cloning, the app will generate custom guided meditations and affirmations based on real-time emotional analysis from user check-ins. This means every calming moment will feel tailor-made.

On the roadmap: multilingual support, mobile apps, AI-generated reflections based on journaling trends, and collaborations with real therapists to bridge the gap between automation and professional guidance.

Built With

• bolt
• clerk
• elevenlabs
• netlify
• next.js
• openai
• shadcn
• supabase
• tailwind