AI-powered emotional wellness journal built for India — write, speak, and understand your emotions.
emoDiary lets users journal their emotions and have real-time voice conversations with an AI companion. Every entry is analysed by Llama (via Groq) to extract multi-label emotion tags and generate a detailed sentiment report. The voice session pipeline uses Groq Whisper for transcription and Sarvam AI for natural Indian-language speech synthesis.
Built for Hindi, Hinglish, Gujarati, and English — because mental health support should speak your language.
- AI Journal — Write freely; AI generates emotion tags, a title, and a full sentiment report
- Voice Therapy Sessions — Real-time voice conversation with a talking avatar (VAD → Whisper → Llama → Sarvam TTS)
- Emotional Insights — Mood trends, emotion distribution charts, and daily streaks
- Therapist Score — AI computes a clinical relevance score; triggers curated therapist recommendation cards with referral code "EmoDiary"
- Multi-language — English, Hindi, Hinglish, Gujarati (switchable mid-session)
- Custom Avatar — Configurable appearance with real-time lip sync during voice sessions
- Free Model — Unlimited journal entries and unlimited voice therapy sessions (no payments)
| Layer | Technology |
|---|---|
| Frontend | Next.js 14, TypeScript, Tailwind CSS, Framer Motion |
| Backend | FastAPI (Python), Uvicorn |
| Database | Supabase (PostgreSQL + Row-Level Security) |
| AI (LLM) | Llama 3 via Groq API |
| AI (STT) | Whisper via Groq API |
| AI (TTS) | Sarvam AI (Indian language voices) |
| Payments | None (completely free) |
| Caching | Upstash Redis |
- Node.js 18+
- Python 3.11+
- A Supabase project
- API keys: Groq, Sarvam AI
cd backend
python -m venv .venv
.venv\Scripts\activate # Windows
# source .venv/bin/activate # Mac/Linux
pip install -r requirements.txtCopy .env.example to .env and fill in your keys:
SUPABASE_URL=
SUPABASE_SERVICE_KEY=
GROQ_API_KEY=
SARVAM_API_KEY=
UPSTASH_REDIS_URL=
UPSTASH_REDIS_TOKEN=
CORS_ORIGINS=http://localhost:3000Run the database migration in Supabase SQL Editor → paste contents of backend/migration.sql → Run.
Start the server:
python -m uvicorn app.main:app --reload --port 8000cd frontend
npm installCreate frontend/.env.local:
NEXT_PUBLIC_SUPABASE_URL=
NEXT_PUBLIC_SUPABASE_ANON_KEY=
NEXT_PUBLIC_API_URL=http://localhost:8000Start the dev server:
npm run devApp runs at http://localhost:3000
| Service | Platform |
|---|---|
| Frontend | Vercel — set root dir to frontend |
| Backend | Render — set root dir to backend, start cmd: uvicorn app.main:app --host 0.0.0.0 --port $PORT |
Set all environment variables on each platform and update CORS_ORIGINS on the backend to your Vercel URL.
| Tier | Price | Limits |
|---|---|---|
| Free | ₹0 | Unlimited (journals + voice sessions) |
emoDiary/
├── backend/
│ ├── app/
│ │ ├── routers/ # journal, chat, voice, subscription, analytics
│ │ ├── services/ # AI, voice, subscription logic
│ │ ├── models/ # DB access + Pydantic schemas
│ │ └── prompts/ # Llama system prompts
│ ├── migration.sql
│ └── requirements.txt
│
├── frontend/
│ └── src/
│ ├── app/
│ │ ├── (protected)/ # dashboard, journal, talk, insights, upgrade
│ │ └── (auth)/ # sign-in, sign-up
│ ├── components/ # avatar, subscription, UI
│ └── hooks/ # useSubscription, etc.
│
└── project_end.md # Full product & business documentation
MIT