VibeBridge

VibeBridge is an emotional AI companion app — a place where users can chat, share their thoughts, and receive empathetic responses

Comment 2

Inspiration

VibeBridge was born from a simple question: "What if there was a place that always listened without judgment?" We believe everyone deserves a space where they feel truly understood — especially when dealing with heavy emotions, anxiety, or loneliness. The core idea was to build an AI that not only "responds" but genuinely "listens" — a companion that is anonymous, but authentic.

What it does

VibeBridge is an emotional AI companion app — a place where users can chat, share their thoughts, and receive empathetic responses. From detecting emotional tones in messages, suggesting coping strategies, to the "Find Your Vibe" feature — which anonymously connects users with others who’ve had similar experiences — everything is designed to create a space for healing that is safe, intelligent, and heartfelt.

How we built it

1. Bolt (Core Infra + AI Persona Chat)

  • Bolt was used to design and deploy the entire AI companion flow: from the first conversation with "Mira" to the point where users receive real human match suggestions.
  • The frontend was built using React and Tailwind CSS, supporting a unified single-window experience for both AI and human chats.
  • We used Bolt to define backend API endpoints for handling messages, scoring engagement, and tracking session state.
  • Bolt also enabled rapid UI development and friendly UX creation from just a few prompts. It also provided convenient connections to GitHub, Netlify, and Supabase, speeding up the full-stack integration process.

2. Supabase (Auth + Realtime Storage)

  • Supabase provides both lightweight authentication and real-time data handling — allowing anonymous users to keep their nickname, emoji, and avatar temporarily without sign-up, while also storing and syncing AI session logs, emotional signals, and match request timestamps seamlessly.
  • Supabase also made backend development simple by offering rich authentication options, easy PostgreSQL database management, and powerful edge functions suitable for building backend APIs. Each user is assigned a unique anonymous ID used for chat with Mira.

3. Netlify (Frontend Hosting)

  • Netlify enables continuous deployment from Git and allowed rapid iteration during hackathon development.

4. AI Engine (Empathetic AI)

  • We used OpenAI GPT-4 with tailored prompts to create "Mira" — a non-robotic, emotionally intelligent persona.
  • A custom-built intent/mood classifier allows the AI to understand when to empathize vs. when to offer guidance.

5. Expo + Mobile App (Mobile MVP)

  • The mobile version was developed using Expo, leveraging shared backend logic via Bolt.
  • Chat UI mirrors the same single-window structure as the web version.
  • Expo also made it fast and developer-friendly to build and publish the mobile app to both the App Store and Google Play.

6. RevenueCat (Monetization System)

  • RevenueCat made it easy to build a premium system: free-tier users have limited daily AI messages and are smoothly guided toward upgrading via a soft paywall.
  • Once the free limit is reached, the app displays a modal: "Ready to stay with Mira a bit longer?"
  • Premium unlocks unlimited chats, custom Mira tone, and priority in match suggestions.

7. Entri + Domain Auto Setup

  • Bolt’s built-in Entri CLI enabled auto-registration and DNS setup for vibebridge.me via IONOS with a single command.
  • HTTPS was configured automatically and linked to our Netlify frontend, minimizing setup time.

Challenges we ran into

  • Tuning the AI to be empathetic without sounding prescriptive.
  • Building a safe, anonymous human-to-human matchmaking system.
  • Maintaining consistent emotional state and tone throughout long chat sessions.

Accomplishments that we're proud of

  • Mira's AI persona truly makes users feel understood.
  • "Find Your Vibe" connects people with similar emotional journeys.
  • We delivered a fully working MVP in a short timeframe thanks to an optimized toolchain.
  • Successfully integrated mobile paywall without disrupting the experience.

What we learned

  • Mental health tech must "feel right," not just "work well."
  • Designing a real-time, emotionally intelligent system requires intentional architecture.
  • Users care less about intelligence — more about sincerity.

What's next for VibeBridge

  • Expanding the matchmaking engine with deeper emotion-based personalization.
  • Introducing long-term features: mood journaling, ritual spaces.
  • Launching VibeBridge globally with multi-language support.
Share this project:

Updates