This is a full-stack AI-powered chat application where users can create and converse with personas powered by Google Gemini. Each persona has its own tone, style, and context, and the app generates intelligent recommendations for how users can improve or continue the conversation — just like having your own personal social coach.
.
├── backend/ # Python FastAPI backend
│ └── requirements.txt
│ └── app.py
├── download/ # React frontend (Next.js)
├── .env # Global environment variables (for backend)
├── .env.local # Frontend (Next.js) environment variables
└── README.md
- Setup virtual environment and install dependencies:
cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
- Run the server:
uvicorn app:main --reload
- Install dependencies:
cd download
npm install
- Run the app:
npm run dev
MONGO_URI=your_mongo_connection_string
MONGO_password=your_mongo_password
GOOGLE_API_KEY=your_gemini_api_key
GOOGLE_CLIENT_ID=your_google_oauth_client_id
GOOGLE_CLIENT_SECRET=your_google_oauth_client_secret
GOOGLE_CLIENT_ID=your_google_oauth_client_id
GOOGLE_CLIENT_SECRET=your_google_oauth_client_secret
NEXTAUTH_URL=http://localhost:3000
NEXTAUTH_SECRET=some_super_secret_key
- Create dynamic AI Personas with unique traits and styles
- Gemini-powered conversation analyzer for:
- Message ratings (1–5)
- Personalized suggestions
- Next-move guidance
- Persistent chat history per session and persona (MongoDB)
- Star-based feedback on user messages
- Google OAuth authentication
- Frontend: React, Next.js, TailwindCSS, NextAuth
- Backend: FastAPI, MongoDB, Google Gemini API
- Database: MongoDB
- Keep
.envand.env.localprivate and never commit them - All chat analytics and recommendations are generated dynamically with Gemini based on persona, history, and user profile
- Support uploading images/memories for personas
- Auto-summarize chats and give weekly reports
- Fine-tuned Gemini or Claude models per persona
- iOS/Android wrapper
Aryan Parab — https://github.com/aryanparab