Conversational Memory Bot with Image Gallery

An intelligent AI-powered application that combines image management with natural language interaction. Built with FastAPI, ChromaDB, and Google's Gemini model, this bot can understand, store, and retrieve images through conversation.

🌟 Features

• Smart Image Gallery

  • Upload and manage images
  • Automatic metadata generation
  • Tag-based organization
  • Newest-first display order

• AI-Powered Search

  • Natural language image search
  • Image similarity search
  • Combined text and image search
  • Semantic understanding

• Conversational Interface

  • Natural language interaction
  • Context-aware responses
  • Image-based discussions
  • Memory of past interactions

• Advanced Image Processing

  • Automatic description generation
  • Smart tagging
  • Vector embeddings
  • Similarity matching

🛠️ Technology Stack

• Backend Framework: FastAPI

• Database: ChromaDB (Vector Database)

• AI Models:

  • Google Gemini 1.5
  • CLIP (ViT-B-32)

• Frontend:

  • HTML5/CSS3
  • JavaScript

• Image Processing: PIL/Pillow

• Vector Operations: NumPy, scikit-learn

• Retrieval-Augmented Generation (RAG) for improved search results powered by LangChain.

📋 Prerequisites

• Python 3.9+
• Google API Key
• 4GB+ RAM
• Storage space for image database

⚙️ Installation

1. Clone the Repository

   git clone [your-repository-url]
   cd memory-bot-ai

2. Create Virtual Environment

   python -m venv venv
   source venv/bin/activate  # Linux/Mac
   # or
   venv\Scripts\activate  # Windows

3. Install Dependencies

   pip install -r requirements.txt

4. Environment Setup Create a .env file in the root directory:

   GOOGLE_API_KEY=your_api_key_here
   ENVIRONMENT=development
   PORT=8000
   HOST=0.0.0.0

🚀 Running the Application

1. Start the Server

   python -m uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

2. Access the Application

   • Main Interface: http://localhost:8000
   • API Documentation: http://localhost:8000/docs

📱 Usage Guide

Image Upload

1. Navigate to the Upload page
2. Select image(s) to upload
3. Wait for automatic processing
4. Review generated descriptions and tags

Gallery Navigation

• Browse images in chronological order
• Use filters and tags for organization
• Click images for detailed view
• Delete unwanted images

Image Search

• Text Search: "Show me images of cats"
• Image Search: Upload a similar image
• Combined Search: Use both text and image

Chat Interface

• Ask about specific images
• Query the image database
• Get image descriptions
• Natural conversation about images

🔧 API Endpoints

Image Management

POST /upload - Upload new images
GET /gallery - Retrieve image gallery
DELETE /image/{id} - Remove an image

Search Operations

POST /search/text - Text-based search
POST /search/image - Image similarity search
POST /search/multimodal - Combined search

Chat Interface

POST /chat - Chat with the bot
GET /chat/history - Retrieve chat history

📁 Project Structure

memory-bot-ai/
├── app/
│   ├── main.py              # FastAPI application entry point
│   ├── config.py            # Configuration settings
│   ├── constants.py         # Global constants and messages
│   ├── services/
│   │   ├── chromadb.py     # Vector database operations
│   │   ├── embeddings.py   # Image/text embeddings
│   │   ├── gemini.py       # AI model integration
│   │   └── llm_langchain.py # Language model chain
│   ├── routes/
│   │   └── app_route.py    # API endpoints
│   ├── templates/
│   │   ├── chat.html       # Chat page
│   │   ├── gallery.html    # Image gallery page
│   │   ├── home.html       # Home page
│   │   └── upload.html     # Image upload page
│   └── static/
│       ├── css/
│          └── style.css   # Main stylesheet
│   ├── chromadb/          # Vector database storage
│   ├── uploads/           # Uploaded image storage
│   └── images/            # Processed images
└── requirements.txt       # Project dependencies

🙏 Acknowledgments

• Google Gemini AI for natural language processing
• OpenAI's CLIP model for image understanding
• ChromaDB for vector storage
• FastAPI community for the excellent framework