Mochi.ai 🧠✨

Mochi.ai is a compassionate, AI-powered mental health companion designed to provide a safe, personalized, and stigma-free space for your thoughts. Built with a modern, secure, and scalable stack, Mochi.ai goes beyond generic chatbots by allowing you to create and interact with unique AI personas.

Core Features

• Secure Authentication: Full user sign-up, sign-in, and management powered by Clerk.
• Persistent Conversations: All chat history is securely saved to your account. No more lost conversations on a page refresh.
• Custom AI Personas: Why talk to a generic bot?
  • Create Your Own: Define a persona's name, description, tone, custom greetings, and even conversational boundaries.
  • Use Pre-mades: Chat with familiar, pre-built characters like Doraemon, Shizuka, or Shinchan, each with a unique personality.
• Mood Logging: Track your mood over time with a simple logging feature.
• Rich AI Responses: The AI provides responses formatted with Markdown (bold, italics) for a more dynamic and expressive chat experience, all powered by the Google Gemini API.

---

💻 Tech Stack

• Frontend: React (Vite), Tailwind CSS
• Backend: FastAPI (Python)
• Database: MongoDB (Atlas) for all user data, personas, and chat history.
• Authentication: Clerk
• AI: Google Gemini API

---

🚀 Getting Started

Follow these instructions to get a copy of the project up and running on your local machine for development and testing.

Prerequisites

You will need to create accounts and get API keys from the following services:

1. Node.js: (v18 or higher)
2. Python: (v3.9 or higher)
3. MongoDB Atlas: A free-tier cluster is sufficient.
   • Get your MongoDB Connection String (URI).
4. Google AI Studio:
   • Get your Gemini API Key.
5. Clerk:
   • Get your Publishable Key (for the frontend).
   • Get your JWKS Endpoint URL (for the backend).

⚙️ Configuration (Environment Variables)

This project has two separate .env files.

1. Backend (backend/.env)

MONGO_URI="your_mongodb_connection_string_here"
GOOGLE_API_KEY="your_gemini_api_key_here"
CLERK_JWKS_URL="https://your-clerk-jwks-endpoint/.well-known/jwks.json"

2. Frontend (frontend/.env)

VITE_CLERK_PUBLISHABLE_KEY="pk_test_...your_publishable_key_here"

Installation & Running

1. Backend Server (FastAPI)

# 1. Navigate to the backend directory
cd backend

# 2. Create a Python virtual environment
python -m venv venv

# 3. Activate the virtual environment
# On Windows:
.\venv\Scripts\activate
# On macOS/Linux:
source venv/bin/activate

# 5. Install all dependencies
pip install -r requirements.txt

# 6. Run the server!
# --reload will automatically restart the server on code changes
uvicorn main:app --reload

2. Frontend (React)

# 1. Navigate to the frontend directory
cd frontend

# 2. Install all npm packages
npm install

# 3. Run the development server
npm run dev