-
-
mindfulvoice
Inspiration
Mental health is deeply personal, yet often inaccessible. While many platforms focus on text journaling, we wanted to build something more natural, emotional, and expressive—just like speaking to a trusted friend. Inspired by therapy sessions, CBT, and self-reflection, MindfulVoice gives users a safe space to talk, feel, and grow—powered by AI and built with Bolt.
What it does
MindfulVoice enables users to record voice journals, receive emotional feedback in real-time, and get personalized CBT-style mental health recommendations. Users interact with a video therapist agent ("Dr. Maya"), track their mood trends over time, and anonymously share their stories in a supportive community hub. The entire experience is designed to be intuitive, private, and emotionally comforting.
How I built it
We built this project entirely on Bolt.new, using its visual UI builder, workflows, and integrations.
- The voice journaling page uses Bolt’s built-in audio interface and workflows to process recordings.
- Voice files are sent to ElevenLabs, which returns emotional tone + voice feedback, played in-app.
- Tavus is embedded as a floating video therapist that responds to user emotions.
- CBT suggestions are displayed using Bolt logic blocks, based on emotion classification.
- A Supabase backend stores user sessions, entries, and community stories anonymously.
- The mood dashboard was designed using Bolt’s chart and layout blocks for insights.
- The app was deployed on Netlify and published on a custom IONOS domain via Entri.
Challenges I ran into
- Voice analysis latency: Reducing the delay between recording and feedback required workflow optimization.
- Tavus sync: Ensuring contextual responses from Dr. Maya based on emotion states was tricky.
- Anonymity vs. insight: We carefully balanced user data privacy with meaningful analytics.
- UI balance: Building an app that felt both clinical and comforting took iterations on color, interaction design, and layout flow.
Accomplishments that I'm proud of
- Built a production-ready emotional health app in Bolt in a short time
- Successfully integrated voice and video AI agents into a low-code platform
- Created a safe space where users can feel heard and understood
- Achieved full eligibility for 5 hackathon challenges with full deployment stack
What I learned
- How to use Bolt.new to build a polished, scalable app with rich features
- The importance of designing emotional UX: voice tone, animations, and empathy matter
- How to stitch together external AI tools (Tavus + ElevenLabs) into a seamless experience
- The unique challenges of building privacy-first mental health products
What’s next for MindfulVoice
- Add multi-language support and adaptive voice tones for diverse users
- Launch a mobile-first progressive web app (PWA)
- Partner with therapists to validate CBT recommendations
- Add mood-based journaling prompts and emotionally intelligent avatars
- Explore integration with Apple Health or Fitbit to enrich mood tracking
Built With
- bolt.new
- boltcharts
- boltuiblocks
- customworkflows
- elevenlabs
- entri
- ionos
- netlify
- responsivelayout
- supabase
- tavus