Octify – AI-Powered Relationship Magic
Connect. Grow. Thrive. 🌟
Both App & App Website build with bolt.new 🌟
💡 Inspiration
Octify was born from a simple question: What if AI could help us navigate the most human thing of all—relationships? Inspired by the emotional complexity of modern connections and the rise of AI-powered wellness tools, we set out to build an app that doesn’t just analyze relationships but transforms them with actionable insights, mood-aware coaching, and immersive role-play.
Our vision? To create a world where understanding your relationships feels less like therapy and more like magic—where every interaction is a step toward deeper connection.
🧠 What It Does
Octify is a native iOS and Android app that helps users:
- Analyze Relationships: Gemini 2.5 Flash generates confidence scores, key insights, and step-by-step strategies for communication issues, trust problems, and more.
- Role-Play with Tavus Avatars: Practice difficult conversations in immersive VR environments (e.g., "I’ll act as your manager — respond how you’d normally react").
- Sync with Wearables: Correlate heart rate variability and sleep data from Apple Watch/Whoop to track physiological stress alongside mood logs.
- Connect with Community: Share anonymized success stories in forums, upvote strategies, and crowdsource solutions for complex relationship dynamics.
🛠️ How We Built It
We built Octify in Bolt.new using:
- Gemini 2.5 Flash API for core insights and voice simulations
- Tavus for AI-generated video avatars in role-play scenarios
- ElevenLabs TTS for mood-aware coaching and accessibility
- Supabase for auth/data storage + Edge Functions to securely connect Gemini
- RevenueCat for premium features (expert Q&A, VR role-play)
We followed 2025 design trends:
- Glassmorphism UI (
rgba(255,255,255,0.1)light mode / dark equivalent) - Calm aesthetics with soft gradients (orange-blue-lavender)
- Ambient animations and haptic feedback
⚠️ Challenges We Ran Into
Prompt Length & Token Management:
- Bolt’s context window strain from long Gemini prompts forced us to adopt modular prompting and
.bolt/ignoreto hide non-essential files (e.g.,node_modules/,dist/). - Solution: Split prompts into screen-by-screen chunks and used Discussion Mode for planning before Build Mode for execution.
- Bolt’s context window strain from long Gemini prompts forced us to adopt modular prompting and
Supabase Integration:
- Mapping persona/challenge data to Supabase tables (
personas,persona_challenges,analysis_results) required meticulous schema design. - Edge Functions became critical for securely passing data to Gemini without exposing raw tokens.
- Mapping persona/challenge data to Supabase tables (
First-Time Mobile Development:
- Transitioning from web to native mobile (via Expo) meant rethinking gestures, animations, and platform-specific theming (Material 3 vs. Apple HIG).
- Gemini occasionally suggested unrealistic strategies, requiring manual tuning of AI responses.
- Transitioning from web to native mobile (via Expo) meant rethinking gestures, animations, and platform-specific theming (Material 3 vs. Apple HIG).
🏆 Accomplishments We’re Proud Of
- ✅ 44 Fully Functional Screens: From splash to VR role-play, all built incrementally in Bolt.new
- ✅ AI-Powered Emotional Intelligence: Gemini’s adaptive tone based on mood tags (😊/😐/😠)
- ✅ Tavus Video Role-Play: Immersive conflict resolution simulations with AI avatars
- ✅ Token-Efficient Workflow: Modular prompts, reusable components (
PersonaCard,ChallengeTag), and.bolt/ignorefor scalability
🧠 What We Learned
- Bolt.new is a game-changer for AI-driven apps, but requires strict token discipline.
- Emotional UX matters: Glassmorphism, ambient animations, and dynamic theming foster trust.
- Cross-platform parity is hard: Material 3 and Apple HIG don’t always align — but it’s worth the effort.
- AI hallucinations happen: Gemini’s occasional unrealistic advice taught us to validate responses with user feedback loops.
🚀 What’s Next for Octify
We’re just getting started. Future plans include:
- Expanding Gemini Integration: Use Gemini 2.5 Pro for advanced expert Q&A
- Deeper Wearable Sync: Integrate biometric data into daily rituals
- Community Gamification: Add badges for top-rated forum strategies
- Voice Journaling: Let users record audio reflections, transcribe with Gemini, and narrate with ElevenLabs
Our goal? To make Octify the Headspace of relationships
"Every conversation is a chance to grow—Octify makes that growth effortless."
Built With
- bolt.new
- edge-functions
- elevenlabs-tts-api
- framer-motion
- google-gemini-2.5-pro/flash-api
- netlify
- postgresql
- react
- revenuecat-web-sdk
- storage)
- supabase-(auth
- tailwind-css
- tavus-ai-video-api
- typescript

Log in or sign up for Devpost to join the conversation.