Calmate - Agentic AI-Powered Mental Health Companion for Corporate Well-being

Calmate is an Agentic AI-powered Mental Health Companion that proactively connects with employees through natural, empathetic conversations on a audio phone call helping them express emotions and release work stress in a private, stigma-free way helping the corporate with anonymized, actionable well-being insights.

🏗️ System Architecture

The project is divided into three main components:

1. Frontend (/frontend):

   • Framework: React 19 with Vite.
   • Styling: TailwindCSS (inferred from design patterns) / CSS Modules.
   • Database Integration: Supabase (Auth & Data).
   • Purpose: Admin dashboard for visualizing data and managing employees.

2. Agentic Backend (/agentic-backed):

   • Framework: FastAPI (Python).
   • Core Logic: Handles webhooks from Vapi.ai, processes call logs, and orchestrates multi-agent analysis using Google Gemini and other LLMs.
   • Database: PostgreSQL (via Supabase).

3. AI Engine (/ai-engine):

   • Framework: Python Scripts.
   • Core Logic: Background workers for deep data analysis, trend calculation, and report generation (PDF).
   • Tools: OpenAI (GPT-4o), Matplotlib (Charts), FPDF (PDF Generation).

🛠️ Tech Stack

• Frontend: React, TypeScript, Vite, Lucide React, Recharts (implied for charts).
• Backend: Python, FastAPI, Uvicorn, Pydantic.
• AI & LLMs: Google Gemini (via Google Generative AI SDK), OpenAI GPT-4o, Vapi.ai (Voice AI).
• Database: PostgreSQL, Supabase.
• DevOps: Docker (potential), Python venv.

🏁 Getting Started

Prerequisites

• Node.js (v18+)
• Python (v3.10+)
• PostgreSQL database (or Supabase project)
• Vapi.ai Account & API Key
• OpenAI API Key
• Google Gemini API Key

1. Clone the Repository

git clone https://github.com/your-username/Calmate-MumbaiHacks.git
cd Calmate-MumbaiHacks

2. Frontend Setup

Navigate to the frontend directory and install dependencies:

cd frontend
npm install

Create a .env file in frontend/ with your Supabase credentials:

VITE_SUPABASE_URL=your_supabase_url
VITE_SUPABASE_ANON_KEY=your_supabase_anon_key

Run the development server:

npm run dev

3. Agentic Backend Setup

Navigate to the backend directory:

cd ../agentic-backed

Create a virtual environment and install dependencies:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Create a .env file in agentic-backed/ with necessary keys:

GOOGLE_API_KEY=your_google_api_key
DATABASE_URL=your_postgres_connection_string

Start the FastAPI server:

uvicorn main:app --reload

4. AI Engine Setup

Navigate to the AI engine directory:

cd ../ai-engine

Install dependencies:

pip install -r requirements.txt

Create a .env file in ai-engine/ with necessary keys:

OPENAI_API_KEY=your_openai_api_key
PG_HOST=your_db_host
PG_PORT=5432
PG_DBNAME=your_db_name
PG_USER=your_db_user
PG_PASSWORD=your_db_password
EMAIL_FROM_ADDRESS=your_email
SMTP_SERVER=smtp.gmail.com
SMTP_PORT=587
SMTP_USER=your_email
SMTP_PASSWORD=your_app_password

Run the worker script manually (or set up a cron job):

python worker.py

📂 Project Structure

Calmate-MumbaiHacks/
├── agentic-backed/       # FastAPI backend for Vapi webhooks & analysis
│   ├── agents/           # AI Agent definitions
│   ├── models/           # Pydantic models
│   ├── main.py           # Entry point
│   └── ...
├── ai-engine/            # Background workers for reporting
│   ├── worker.py         # Main reporting script
│   └── ...
├── frontend/             # React Admin Dashboard
│   ├── src/
│   ├── package.json
│   └── ...
└── README.md             # Project documentation

🤝 Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements.

📄 License

This project is licensed under the MIT License.