Highlights
Project snapshot
- Built around Solace, a contextual AI companion that combines memory, emotional detection, and guided support.
- Uses Expo, Supabase, and OpenAI orchestration to deliver streaming, multi-service mobile AI experiences.
- Blends emotional analytics, voice interaction, social sharing, subscriptions, and native mobile features in one product.
Overview
A mobile-first wellness system, not a passive mood tracker
Feel is an AI emotional wellness platform built around Solace, an intelligent conversational companion designed for personalized, context-aware support.
The product treats AI as an orchestration layer across mood detection, memory retrieval, content, actions, and voice rather than as a simple chat interface.
Problem
Most wellness apps are passive and forgetful
- Traditional mood trackers collect data without offering meaningful, contextual support.
- Many chatbots lack memory and personalization, so conversations feel generic and repetitive.
- Users need structured emotional support systems that connect reflection to action.
Solution
Feel combines emotional intelligence, memory, and guided action
- Designed a multi-layer system for emotional detection, memory retrieval, curated content, and real-world action suggestions.
- Used streaming AI and tool-based orchestration so responses can react to user context instead of returning static advice.
- Added voice, analytics, and social engagement systems to increase accessibility and long-term retention.
Core Features
Support loops designed for personalization and retention
Solace AI Companion
- Streaming AI chat
- Tool-based orchestration
- Context-aware responses
Mood and Memory Systems
- Multi-emotion detection with confidence scoring
- Mood tracking and shift analysis
- Embeddings and conversation history for context reconstruction
Personalized Support
- Curated content and guided practices
- Action recommendations
- Emotional analytics, streaks, and behavioral pattern views
Voice, Social, and Monetization Layers
- Whisper transcription and voice onboarding
- Invite system and mood sharing
- RevenueCat subscriptions with free-tier limits
Native and Experimental Features
- iOS widgets, haptics, and deep linking
- Feature-flagged Solana and Arweave data ownership layer
Architecture
Mobile client with AI orchestration and serverless data layers
Expo app
→ Vercel AI orchestrator
→ Supabase Edge Functions
→ Supabase DB
Technical Highlights
AI orchestration is treated as product infrastructure
- Built around real-time streaming AI instead of request-response chat only.
- Uses a multi-service architecture with memory, emotion analysis, and support delivery as separate concerns.
- Combines React Native product work with backend systems for auth, subscriptions, and analytics-driven engagement loops.
Impact
A scalable foundation for deeply personalized emotional support
- Creates stronger personalization through memory-aware conversations and emotional context reconstruction.
- Builds engagement loops through streaks, insights, voice interaction, and actionable guidance.
- Establishes a scalable wellness platform with room for native, social, and experimental data-ownership features.