Designed for ADHD, autistic, dyslexic, and neurodivergent users
Lower cognitive load • Voice-first interface • Intelligent scheduling • Compassionate support
FlowMind is a therapeutic planning companion that uses AI-driven conversation and schedule intelligence to help neurodivergent users manage their day with less stress. Powered by IBM NeuralSeek's mAIstro for sentiment analysis and OpenAI Whisper for voice transcription.
Watch the full demo: https://www.youtube.com/watch?v=4L5e3T1ukhQ
- 🎙️ Voice Mood Check-ins: Speak naturally about how you're feeling
- 🤖 AI Sentiment Analysis: mAIstro analyzes mood with schedule correlation
- 🔊 TTS Responses: ElevenLabs generates empathetic voice responses
- 📊 Mood Tracking: Continuous mood patterns with schedule impact analysis
- 💬 Multi-turn Conversations: Natural dialogue that remembers context
- 📅 Google Calendar Sync: Automatic schedule analysis
- ⏰ Schedule Intensity Detection: High/medium/low workload calculation
- 🎯 Agentic Activity Generation: AI fills gaps with workouts, meals, breathing breaks
- 🔄 Real-time Updates: Webhook-based calendar monitoring
- 📈 Pattern Discovery: AI learns optimal activity timing from your habits
- Calm UI Mode: Reduced motion, high contrast (WCAG AAA), larger touch targets
- Micro-Steps: Every task broken into 3-5 concrete actions
- Choice Architecture: Max 2 options to reduce decision fatigue
- Predictable Timing: 10-3-1 minute nudges (gentle haptics)
- Energy-Aware: Activities only during your peak energy windows
- No Guilt Design: Skip buttons always visible, no streak shaming
- Local-First: Profile stored securely on device (expo-secure-store)
- Supabase Backend: Optional cloud sync with row-level security
- OAuth 2.0: Secure Google Calendar integration via Auth0
- Encrypted Storage: All sensitive data encrypted at rest
┌──────────────────────────────────────────────────────────┐
│ iOS App (Expo/React Native) │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐│
│ │ Today │ │ Explore │ │ Mood │ │ Profile ││
│ │ View │ │ Tab │ │ Check-in │ │ Settings ││
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘│
│ ↓ ↓ ↓ ↓ │
│ ┌────────────────────────────────────────────────────┐ │
│ │ Voice Recording (expo-av) + Secure Storage │ │
│ └────────────────────────────────────────────────────┘ │
└────────────────────────┬─────────────────────────────────┘
│ HTTPS/REST API
▼
┌──────────────────────────────────────────────────────────┐
│ Node.js Backend (Express) │
│ ┌─────────────────┐ ┌─────────────────┐ │
│ │ API Routes │ │ Services │ │
│ │ • Users │ │ • Conversational│ │
│ │ • Conversation │ │ Agent │ │
│ │ • Calendar │ │ • Calendar Sync│ │
│ │ • Mood │ │ • Agentic │ │
│ │ • Orchestration│ │ • Health │ │
│ └─────────────────┘ └─────────────────┘ │
└────┬────────┬──────────┬─────────┬─────────────────────┘
│ │ │ │
▼ ▼ ▼ ▼
┌─────────┐ ┌──────────┐ ┌─────────┐ ┌─────────────────┐
│Supabase │ │NeuralSeek│ │ OpenAI │ │ ElevenLabs │
│PostgreSQL│ │ mAIstro │ │ Whisper │ │ TTS (Rachel) │
│ │ │ │ │ │ │ │
│ • Users │ │ • Sentiment│ │ • Voice │ │ • Voice Response│
│ • Moods │ │ Analysis│ │ Transcr│ │ • Calm tone │
│ • Plans │ │ • Schedule│ │ iption │ │ • Empathetic │
│ • Events│ │ Correlation│ │ │ │ │
└─────────┘ └──────────┘ └─────────┘ └─────────────────┘
┌──────────────────────────────────┐
│ Google Calendar API │
│ • OAuth 2.0 via Auth0 │
│ • Webhook subscriptions │
│ • FreeBusy queries │
│ • Event creation │
└──────────────────────────────────┘
- Expo SDK 52 - Cross-platform mobile framework
- TypeScript - Type safety and developer experience
- expo-av - Audio recording for voice input
- expo-haptics - Sensory-friendly tactile feedback
- expo-secure-store - Encrypted local profile storage
- @react-native-google-signin/google-signin - Google OAuth
- @supabase/supabase-js - Cloud database client
- Express 4.x - REST API framework
- ES Modules - Modern JavaScript syntax
- Supabase Client - PostgreSQL database access
- node-fetch - HTTP client for external APIs
- Webhook System - Google Calendar real-time updates
- IBM NeuralSeek mAIstro - Agentic Framework Orchestration to extract AI sentiment analysis & conversational responses
- OpenAI Whisper - Voice-to-text transcription ($0.006/min)
- ElevenLabs TTS - Text-to-speech with Rachel voice (10k chars/month free)
- Google Calendar API - Schedule synchronization & event management
- Supabase - PostgreSQL database with real-time subscriptions
- Auth0 - OAuth 2.0 authentication provider
Core Tables:
-- User management (email-based, Auth0 optional)
users (id UUID, email, name, auth0_sub, created_at)
-- Neurodivergent preferences & personality
user_profiles (user_id, neuro_preferences JSONB, personality_traits JSONB)
-- Voice mood check-ins with AI analysis
mood_check_ins (id, user_id, transcription, mood_score, ai_analysis JSONB)
-- Weekly schedule context
weekly_schedules (id, user_id, avg_daily_density, daily_breakdown JSONB)
-- AI-discovered patterns
mood_patterns (id, user_id, pattern_type, confidence_score, recommendations JSONB)
-- Conversation history for mAIstro context
conversations (id, user_id, role, message, mood_score, created_at)
-- Calendar sync events
calendar_sync_events (id, user_id, event_id, start_time, end_time, title)
-- AI-generated activities
agentic_activities (id, user_id, activity_type, title, start_time, duration_sec)- Node.js 18+ and npm
- Xcode 14+ (for iOS development)
- Expo CLI (
npm install -g expo-cli) - Accounts Required:
- Supabase (free tier: 500MB database, 1GB file storage)
- IBM NeuralSeek (free trial available)
- OpenAI ($5 free credit for Whisper API)
- ElevenLabs (10,000 characters/month free)
- Google Cloud Console (Calendar API)
- Auth0 (free: 7k users, social login)
git clone https://github.com/willofcode/flowmind.git
cd flowmind
# Install server dependencies
cd server
npm install
# Install client dependencies
cd ../client
npm install- Create a Supabase project
- Run the complete schema:
# Copy SQL from supabase-schema.sql in root directory # Paste into Supabase SQL Editor and execute
- Copy your project URL and anon key
Follow detailed guides in:
Guide/GOOGLE_CALENDAR_INTEGRATION_COMPLETE.md- Complete OAuth setupGuide/QUICK_TEST_GUIDE.md- Testing checklist
Quick Steps:
- Enable Calendar API in Google Cloud Console
- Create OAuth 2.0 credentials (iOS + Web app)
- Add authorized redirect URIs:
https://dev-YOUR_TENANT.us.auth0.com/login/callbackcom.googleusercontent.apps.YOUR_CLIENT_ID:/oauth2redirect/google
- Download
GoogleService-Info.plist→client/folder - Set up ngrok tunnel for webhooks:
ngrok http 3001 # Copy HTTPS URL to GOOGLE_CALENDAR_WEBHOOK_URL
See Guide/VOICE_TRANSCRIPTION_ELEVENLABS_SETUP.md for complete guide.
Quick OpenAI Whisper Setup:
- Get API key from OpenAI Platform
- Add to
client/.env:EXPO_PUBLIC_OPENAI_API_KEY=sk-your_key_here
- Cost: $0.006 per minute ($5 credit = ~833 minutes)
ElevenLabs TTS Setup:
- Get API key from ElevenLabs
- Add to
server/.env:ELEVENLABS_API_KEY=sk_your_key_here
- Free tier: 10,000 characters/month (~20-30 conversations)
Terminal 1 - Backend Server:
cd server
npm start
# Should see:
# ✅ Database connected
# ✅ NeuralSeek connected
# 🎉 FlowMind API Server running on http://localhost:3001Terminal 2 - iOS Client:
cd client
npx expo prebuild # First time only - generates native iOS folder
npm run ios # Opens Xcode simulator-
Mood Check-in with Voice:
- Open app → Explore tab
- Tap conversation bubble icon
- Tap microphone 🎤
- Speak: "I'm feeling great today!"
- Watch AI analyze sentiment + play TTS response
-
Calendar Sync:
- Sign in with Google
- App automatically fetches schedule
- AI generates activities based on free time
-
Breathing Sessions:
- Navigate to calm session screen
- Choose 5/10/15 minute session
- Listen to guided TTS breathing exercise
- Open app → Profile tab
- Answer questions about:
- Activities: Preferred workouts (walks, yoga, weights...)
- Energy Windows: When you have focus/energy (e.g., 10am-12pm, 4pm-6pm)
- Focus Limits: Maximum focus block length (e.g., 25 minutes)
- Sensory Preferences:
- Reduced motion animations
- Haptics-only feedback (no sounds)
- High contrast mode
- Silent mode schedule
- Diet: Style (Mediterranean, vegan...) & restrictions
- Sleep: Usual bedtime & wake time
- Buffer Policy: Pre/post activity padding (default: 10 min)
- Explore Tab → Tap floating conversation icon
- Record Voice Message:
- Tap microphone 🎤
- Speak naturally: "I'm feeling overwhelmed with today's schedule"
- Tap stop when done
- AI Analysis:
- OpenAI Whisper transcribes your voice
- mAIstro analyzes sentiment + correlates with calendar
- Mood score calculated (1-10)
- Schedule impact assessed (high/medium/low intensity)
- Empathetic Response:
- AI generates supportive response
- ElevenLabs TTS plays voice response (Rachel voice)
- Recommendations displayed
- Continue Conversation:
- Multi-turn dialogue supported
- AI remembers context from previous turns
- Can type text or use voice
- Today tab shows your next scheduled activity
- See 3 micro-steps with concrete actions:
- ✅ "Fill water bottle"
- ✅ "Put on sneakers"
- ✅ "Go outside for 5-minute walk"
- Tap "Start" when ready
- Get 10-3-1 minute nudges (gentle haptic feedback)
- Complete micro-steps one by one
- Tap "Complete" when done
- Options:
- Swap Alternative: Quick A/B alternative activity
- Skip: No guilt, always available
- Google Calendar Sync:
- App fetches your schedule every 10 seconds
- Calculates schedule intensity:
- High (>70% busy): Insert 5-10 min breathing breaks only
- Medium (40-70%): Add movement snacks + meals
- Low (<40%): Full workouts + meal prep + optional activities
- Automatic Gap Filling:
- AI detects free time windows
- Generates appropriate activities
- Respects energy windows from profile
- Applies buffer rules (10 min before/after)
- Activity Types:
- 🏃 Workouts (15-60 min)
- 🍽️ Meals (30-45 min)
- 🧘 Breathing sessions (5-15 min)
- 🚶 Movement snacks (10 min)
- Navigate to Calm Session screen
- Choose duration: 5 / 10 / 15 minutes
- AI generates personalized breathing script
- ElevenLabs creates soothing audio guidance
- Play session:
- Calm voice (Rachel)
- Gentle pacing
- Grounding techniques
- Can pause/resume
- Sessions cached for reuse (saves API costs)
flowmind/
├── client/ # Expo/React Native iOS app
│ ├── app/ # App screens (Expo Router)
│ │ ├── (tabs)/ # Tab navigation
│ │ │ ├── index.tsx # Today View - next activity
│ │ │ ├── explore.tsx # Browse + conversation mode
│ │ │ ├── mood-insights.tsx # Mood tracking history
│ │ │ └── profile.tsx # Settings & preferences
│ │ ├── welcome.tsx # Conversation mode (voice/text)
│ │ ├── mood-checkin.tsx # Voice mood check-in
│ │ ├── calm-session.tsx # Breathing/meditation TTS
│ │ ├── sign-in.tsx # Google OAuth login
│ │ └── _layout.tsx # App navigation structure
│ ├── components/ # Reusable UI components
│ │ ├── today-view.tsx # Main focus card with micro-steps
│ │ ├── task-bubble.tsx # Activity cards with shapes
│ │ ├── mood-checkin-stt.tsx # Voice recording component
│ │ ├── conversational-mood-checkin.tsx # Multi-turn conversation
│ │ └── ui/ # Base components (button, card, etc.)
│ ├── lib/ # Client-side logic
│ │ ├── api-client.ts # Backend API wrapper
│ │ ├── profile-store.ts # Encrypted local storage
│ │ ├── google-auth.ts # Google OAuth client
│ │ ├── voice-transcription.ts # OpenAI Whisper integration
│ │ └── notification-manager.ts # 10-3-1 nudge system
│ ├── types/ # TypeScript definitions
│ │ └── neuro-profile.ts # Core data types
│ ├── constants/ # Design tokens
│ │ ├── calm-theme.ts # Neurodivergent-friendly design
│ │ └── theme.ts # Original theme
│ ├── GoogleService-Info.plist # Google OAuth config (iOS)
│ └── app.json # Expo configuration
├── server/ # Node.js/Express backend
│ ├── src/
│ │ ├── routes/ # API endpoints
│ │ │ ├── conversation.routes.js # Voice mood analysis
│ │ │ ├── calendar-sync.routes.js # Google Calendar webhooks
│ │ │ ├── agentic.routes.js # AI activity generation
│ │ │ ├── mood.routes.js # Mood tracking
│ │ │ ├── users.routes.js # User management
│ │ │ ├── schedules.routes.js # Weekly schedules
│ │ │ ├── orchestration.routes.js # AI orchestration
│ │ │ └── health.routes.js # Health check
│ │ ├── services/ # Business logic
│ │ │ ├── conversational-agent.service.js # mAIstro + TTS
│ │ │ ├── calendar-sync.service.js # Google Calendar
│ │ │ └── agentic-activities.service.js # Activity generation
│ │ ├── config/ # Configuration
│ │ ├── middleware/ # Express middleware
│ │ └── utils/ # Helper functions
│ ├── index.js # Server entry point
│ ├── .env.example # Environment variable template
│ └── package.json
├── Guide/ # Implementation documentation
│ ├── CONVERSATION_INTEGRATION_COMPLETE.md # Voice conversation guide
│ ├── GOOGLE_CALENDAR_INTEGRATION_COMPLETE.md # OAuth setup
│ ├── VOICE_TRANSCRIPTION_ELEVENLABS_SETUP.md # STT/TTS guide
│ ├── CALENDAR_OPTIMIZER_ARCHITECTURE.md # AI scheduling logic
│ ├── DESIGN_PATTERNS.md # UX principles
│ └── *.md # Additional guides
├── supabase-schema.sql # Complete database schema
├── README.md # This file
└── .gitignore # Excluded files
- One thing at a time - Today view shows ONLY next block
- Minimal choices - Max 2 options (A/B)
- Clear hierarchy - Large titles, obvious CTAs
- High contrast - AAA WCAG compliance
- Reduced motion - Optional animations
- Haptics over sound - Non-intrusive feedback
- Larger touch targets - 48-56px minimum
- Consistent timing - Same windows daily
- Buffers - Always 10 min before/after
- Fallbacks - Never "no plan" - always micro-option
- No guilt - Skip button always visible
- Gentle nudges - "Put on shoes" not "WORKOUT NOW!"
- Energy-aware - Never push when you're depleted
- Expo Speech (TTS) for micro-step narration
- Voice commands via Whisper API
- Hands-free mode toggle
- Adaptive nudge timing based on response patterns
- Custom nudge messages per user
- Silent mode schedule (e.g., silent after 9pm)
- Real-time STT (no stop button needed)
- Emotion detection from voice tone
- Multi-language support (Spanish, French, German)
- Offline voice processing (on-device ML)
- Voice-only mode (fully hands-free)
- Android version (React Native)
- Web dashboard (insights & history)
- Apple Watch complications
- Siri shortcuts integration
- Calendar widget for iOS home screen
Voice Features:
- Record voice mood check-in (Explore tab → mic button)
- Verify Whisper transcription accuracy
- Check mAIstro sentiment analysis in console logs
- Confirm TTS audio plays with Rachel voice
- Test multi-turn conversation (follow-up questions)
- Verify mood score calculation (1-10 scale)
Calendar Integration:
- Sign in with Google account
- Confirm calendar events appear in app
- Verify schedule intensity calculation
- Check AI-generated activities in gaps
- Test webhook updates (real-time sync)
UI/UX:
- Toggle reduced motion (Profile → Sensory)
- Test haptic feedback on all buttons
- Verify high contrast mode
- Check touch target sizes (≥48px)
- Test micro-steps display (3-5 items)
Backend Health:
curl http://localhost:3001/health
# Expected response:
{
"status": "healthy",
"database": "connected",
"neuralseek": "connected",
"elevenlabs": "configured"
}We welcome contributions, especially from neurodivergent developers!
- Read
Guide/DESIGN_PATTERNS.mdfor UX principles - Check GitHub Issues for tasks
- Fork the repo and create a feature branch
- Follow existing code style (TypeScript, ESLint)
- Test on real iOS devices when possible
- Submit PR with clear description
- Android support (React Native)
- Offline mode improvements
- Voice command expansion
- Localization (i18n)
- Performance optimization
- Accessibility improvements
MIT License - See LICENSE file
Copyright (c) 2025 FlowMind Team
Built with love for the neurodivergent community at SBU Hack25'.
Special Thanks:
- ADHD & autistic beta testers who provided invaluable feedback
- Accessibility advocates who reviewed our design principles
- IBM NeuralSeek team for mAIstro support
- ElevenLabs for calming TTS voices
- OpenAI for Whisper transcription accuracy
- Neurodivergent design experts and consultants
Inspiration:
This app was built by someone with ADHD, for people who think differently. Every feature is designed to reduce friction, not add it.
- GitHub Issues: Report bugs or request features
- GitHub Discussions: Ask questions or share ideas
- Documentation: See
Guide/folder for detailed guides - Email: [email protected] (coming soon)
- Voice Transcription Setup
- Google Calendar Integration
- Design Patterns
- Conversation Mode
- Calendar Optimizer
Made with 🧠 for minds that flow differently
"The world wasn't built for us. So we built our own tool."