GenAI Playground

A curated collection of Generative AI applications, rapid prototypes, and learning experiments built while exploring modern AI technologies.

This repository contains smaller AI projects that focus on experimenting with Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), AI Agents, Prompt Engineering, and Full-Stack AI Development.

Unlike flagship repositories, these projects are intentionally lightweight and exploratory — serving as proof-of-concepts, learning exercises, and rapid prototypes.

Technologies

Category	Technologies
Languages	Python, HTML, CSS, JavaScript
LLMs	Google Gemini 2.0 Flash, Ollama (Llama 3.2), Groq (LLaMA 3 8B)
Frameworks	Flask, Streamlit
AI/ML Libraries	LangChain, LangChain-Google-GenAI, google-generativeai, Ollama Python SDK, Sentence Transformers, HuggingFace Transformers
Vector Databases	ChromaDB, LangChain-Chroma
NLP / Embeddings	Sentence Transformers (MiniLM-L6-v2, paraphrase-MiniLM-L6-v2), HuggingFace Embeddings
Speech / Voice	OpenAI Whisper, SpeechRecognition, ElevenLabs TTS, Google Speech Recognition, PyAudio
Orchestration	LangChain (Chains, Prompt Templates, Memory, Text Splitters)
Data & Visualization	Pandas, NumPy, Matplotlib, Seaborn, SQLite
External APIs	World Time API, Open Exchange Rates API, Groq API, ElevenLabs API
Security	OWASP Top 10 2021, Automated vulnerability analysis
Frontend	Streamlit (rapid prototyping), Custom HTML/CSS/JS (production-style UI), Font Awesome, Animate.css
Environment	python-dotenv, virtualenv, Flask-CORS
Templating	Jinja2 (Flask templates)

Projects

AI Data Science Tutor

An interactive AI-powered Data Science tutor that provides expert guidance, explanations, and code examples for various data science concepts.

Detailed Documentation: Data Science Tutor AI — Comprehensive Documentation

Features

Interactive Q&A — Ask any data science question and get detailed, expert responses
Code Examples — Receive Python code examples for practical implementation
Conversational Memory — Version 1 maintains chat history for contextual, multi-turn responses
Wide Coverage — Supporting topics including:
- Machine Learning Algorithms
- Statistical Analysis
- Data Preprocessing
- Python Data Science Libraries (NumPy, Pandas, Scikit-learn, etc.)
- Data Visualization
- Model Evaluation
- Advanced Data Science Techniques

Version 1 — Streamlit + Gemini + LangChain

Area	Details
Frontend	Streamlit
LLM	Google Gemini 2.0 Flash
Orchestration	LangChain (LLMChain, ChatPromptTemplate, ConversationBufferMemory)
Key Feature	Conversational memory for multi-turn context

Version 2 — Flask + Ollama (Self-Built)

Area	Details
Frontend	Custom HTML/CSS
Backend	Flask
LLM	Ollama (Llama 3.2, locally hosted)
Key Feature	Fully offline, custom UI, enhanced prompt engineering

AI Code Reviewer

A web-based AI-powered tool that reviews Python code, detects bugs, suggests improvements, and generates human-friendly documentation using language models.

Detailed Documentation: AI Code Reviewer — Documentation

Features

Bug Detection — Identify potential bugs and errors in Python code
Code Improvement — Get suggestions for better code quality, readability, and structure
Corrected Code — Receive fixed versions of submitted code
Documentation Generation — Generate human-friendly explanations of code functionality
Context Filtering — Strict prompt rules to only respond to code-related queries

Version 1 — Streamlit + Gemini API

Area	Details
Frontend	Streamlit
LLM	Google Gemini 2.0 Flash (via `google.generativeai`)
Config	`.env` file for API key management
Key Feature	Quick prototyping with simple input/output layout

Version 2 — Flask + Ollama (Self-Built)

Area	Details
Frontend	Custom HTML/CSS
Backend	Flask
LLM	Ollama (Llama 3.2, locally hosted)
Key Feature	Offline processing, robust prompt engineering, custom UI

Advanced Semantic Search Engine

An advanced semantic search engine that identifies movies based on memorable quotes and dialogue lines using Natural Language Processing and vector embeddings. Includes a Marvel Cinematic Universe subtitle dataset with 23 movies.

Detailed Documentation: Building a Movie Subtitle Search Engine — From AI-Assisted to Self-Coded Development

Features

Semantic Search — Find movies using natural language queries, not just exact keyword matches
Vector Embeddings — Uses Sentence Transformers for accurate semantic matching
Real-time Search — Fast similarity search with confidence scoring
Audio Support — Convert speech to text for voice-based searches (Version 1)
Advanced Filtering — Filter by genre, year, language, and sentiment (Version 1)
Data Visualization — Sentiment distribution and movie analytics (Version 1)
RESTful API — JSON endpoints for search and health check (Version 2)
Extensible — Easy to add new movie subtitle files and expand the database

Version 1 — Python + Sentence Transformers + Whisper (AI-Assisted)

Area	Details
Frontend	CLI / Script-based
Embeddings	Sentence Transformers (`all-MiniLM-L6-v2`)
Vector DB	ChromaDB
Database	SQLite (in-memory)
Speech-to-Text	OpenAI Whisper
Visualization	Matplotlib, Seaborn
Key Feature	Advanced analytics, audio processing, sentiment filtering

Version 2 — Flask + LangChain + ChromaDB (Self-Built)

Area	Details
Frontend	Custom HTML/CSS/JS
Backend	Flask
Embeddings	HuggingFace Embeddings (`paraphrase-MiniLM-L6-v2`)
Vector DB	ChromaDB (via LangChain-Chroma)
Text Processing	LangChain CharacterTextSplitter
Key Feature	Clean web UI, RESTful API, efficient chunked search

OWASP AI Vulnerability Scanner

AI-powered OWASP Top 10 vulnerability scanner using Ollama and LLaMA for automated code and web security analysis with a Flask interface. Supports both direct code input and URL-based content fetching.

Detailed Documentation: OWASP AI Vulnerability Scanner

Features

AI-Powered Analysis — Uses LLaMA 3.2 model via Ollama for intelligent vulnerability detection
OWASP Top 10 Coverage — Comprehensive analysis covering all OWASP Top 10 2021 categories (A01–A10)
Dual Input Methods — Analyze source code directly or fetch content from URLs
Risk Assessment — Categorizes vulnerabilities by severity (Critical, High, Medium, Low)
Detailed Reports — Provides vulnerability descriptions, locations, and remediation suggestions
RESTful API — JSON API endpoints (/analyze, /health) for integration with other tools
Interactive Web Interface — Modern, responsive UI with tabbed input, loading states, and expandable results

Area	Details
Frontend	Custom HTML/CSS/JS (inline template)
Backend	Flask
LLM	Ollama (Llama 3.2, locally hosted)
Security Framework	OWASP Top 10 2021
API	RESTful JSON (`/analyze`, `/health`)
Key Feature	Dual code/URL analysis, severity-based risk badges, structured JSON reports

AI Travel Planner

An AI-powered travel recommendation app that combines Google Gemini and Ollama to provide personalized destination insights, including restaurants, hotels, attractions, real-time weather, local time, and currency exchange rates.

Detailed Documentation: AI Travel Planner — Technical Documentation

Features

Travel Recommendations — AI-generated suggestions for restaurants, hotels, attractions, and activities
Local Information — Current weather, local time, and currency exchange rates for any destination
Destination Insights — Cultural and contextual information powered by Gemini
Multi-Model Pipeline — Gemini provides context, Ollama generates personalized recommendations
Modern UI — Dark-themed, responsive interface with smooth animations (Version 2)
Safety Tips — Travel safety recommendations and precautions (Version 1)
Weather API — Dedicated endpoint for location-based weather data

Version 1 — Streamlit + Gemini + LangChain (AI-Assisted)

Area	Details
Frontend	Streamlit (dark-themed custom CSS)
LLM	Google Gemini 2.0 Flash
Orchestration	LangChain (LLMChain, PromptTemplate)
Key Feature	Travel options with cost estimates, safety tips, multi-mode transport

Version 2 — Flask + Gemini + Ollama (Self-Built)

Area	Details
Frontend	Custom HTML/CSS/JS (Font Awesome, Animate.css)
Backend	Flask
LLMs	Google Gemini 2.0 Flash + Ollama (Llama 3.2)
External APIs	World Time API, Open Exchange Rates API
Key Feature	Multi-model pipeline, real-time weather/currency, RESTful weather endpoint

Medical Receptionist Voice Agent

An AI-powered medical receptionist system that handles patient verification calls through both text and voice interactions. Features voice-to-voice conversation with multiple AI voice options, powered by Groq LLaMA 3, ElevenLabs TTS, and Google Speech Recognition.

Features

Voice-to-Voice Interaction — Full conversational AI with speech recognition and text-to-speech
Patient Verification — Automated verification against a patient database using name, phone, and date of birth
Multiple Voice Options — 8 different AI voices (male/female, various tones) via ElevenLabs
Session Management — Maintains conversation context across interactions
Real-time Audio — Instant speech synthesis and playback
RESTful API — Clean endpoints for text chat, voice chat, TTS, audio retrieval, and session reset
Health Check — Status endpoint with active session count and feature flags

Area	Details
Frontend	Custom HTML/JS (template-based)
Backend	Flask, Flask-CORS
LLM	Groq LLaMA 3 8B (via Groq API)
Text-to-Speech	ElevenLabs API (multilingual v2, 8 voice options)
Speech-to-Text	Google Speech Recognition (via SpeechRecognition + PyAudio)
Data	JSON-based patient records
API	RESTful (`/chat`, `/voice_chat`, `/speak`, `/get_audio`, `/voices`, `/status`, `/reset`)
Key Feature	Voice-to-voice medical receptionist, multi-voice selection, session-based conversation

Repository Structure

GenAI-Playground/
│
├── AI_DS_Tutor/                        # Data Science Tutor AI
│   ├── version_1/
│   │   └── AI_DS_Tutor.py              # Streamlit + LangChain + Gemini app
│   ├── version_2/
│   │   ├── app.py                      # Flask application entry point
│   │   ├── utils/
│   │   │   └── functions.py            # Ollama integration & prompt logic
│   │   └── templates/
│   │       └── index.html              # Custom HTML/CSS frontend
│   ├── requirements.txt                # Combined dependencies
│   └── README.md                       # Project-specific documentation
│
├── AI_code_reviewer/                   # AI Code Reviewer
│   ├── version_1/
│   │   ├── app.py                      # Streamlit + Gemini API app
│   │   └── .env                        # Google API key config
│   ├── version_2/
│   │   ├── app.py                      # Flask application entry point
│   │   ├── utils/
│   │   │   └── functions.py            # Ollama integration & review logic
│   │   └── templates/
│   │       └── index.html              # Custom HTML/CSS frontend
│   └── README.md                       # Project-specific documentation
│
├── Advanced-Semantic-Search-Engine/    # Movie Subtitle Search Engine
│   ├── version_1/
│   │   └── AI_SEO.py                   # Sentence Transformers + Whisper + ChromaDB
│   ├── version_2/
│   │   ├── app.py                      # Flask + LangChain-Chroma app
│   │   ├── templates/
│   │   │   └── index.html              # Custom HTML/CSS/JS frontend
│   │   ├── subtitles/                  # Movie subtitle files (.txt) — 23 MCU films
│   │   └── app.ipynb                   # Jupyter notebook demo
│   ├── requirements.txt                # Combined dependencies
│   └── README.md                       # Project-specific documentation
│
├── owasp-ai-vulnerability-scanner/     # OWASP AI Vulnerability Scanner
│   ├── app.py                          # Flask app with inline HTML template
│   └── README.md                       # Project-specific documentation
│
├── Travel_planner/                     # AI Travel Planner
│   ├── version_1/
│   │   └── Travel_Planner.py           # Streamlit + LangChain + Gemini app
│   ├── version_2/
│   │   ├── app.py                      # Flask + Gemini + Ollama app
│   │   ├── .env                        # Gemini API key config
│   │   └── templates/
│   │       └── index.html              # Dark-themed HTML/CSS/JS frontend
│   ├── requirements.txt                # Combined dependencies
│   └── README.md                       # Project-specific documentation
│
├── Voice_agents/                       # Medical Receptionist Voice Agent
│   ├── main.py                         # Flask + Groq + ElevenLabs app
│   ├── templates/
│   │   └── index.html                  # Web interface for chat & voice
│   └── README.md                       # Project-specific documentation
│
└── README.md                           # ← You are here

Getting Started

Prerequisites

Python 3.8+
Google API Key (for Streamlit/Gemini versions) — Get one here
Ollama (for Flask/local LLM versions) — Download here
Groq API Key (for Voice Agent) — Get one here
ElevenLabs API Key (for Voice Agent TTS) — Get one here

Clone the Repository

git clone https://github.com/MONARCH1108/GenAI-Playground.git
cd GenAI-Playground

AI Data Science Tutor

Version 1 — Streamlit + Gemini

# Install dependencies
pip install streamlit langchain langchain-google-genai

# Set your Google API key
export GOOGLE_API_KEY="your-api-key"          # Linux/Mac
set GOOGLE_API_KEY=your-api-key               # Windows CMD
$env:GOOGLE_API_KEY="your-api-key"            # Windows PowerShell

# Run the app
streamlit run AI_DS_Tutor/version_1/AI_DS_Tutor.py

Version 2 — Flask + Ollama

# Install dependencies
pip install flask ollama

# Pull the Llama 3.2 model (requires Ollama running)
ollama pull llama3.2:latest

# Run the app
python AI_DS_Tutor/version_2/app.py

The application will be available at http://127.0.0.1:5000/

AI Code Reviewer

Version 1 — Streamlit + Gemini

# Install dependencies
pip install streamlit google-generativeai python-dotenv

# Set your Google API Key in .env
echo GOOGLE_API_KEY=your_key_here > AI_code_reviewer/version_1/.env

# Run the app
streamlit run AI_code_reviewer/version_1/app.py

Version 2 — Flask + Ollama

# Install dependencies
pip install flask ollama

# Make sure Ollama is installed and the model is pulled
ollama pull llama3.2:latest

# Run the app
python AI_code_reviewer/version_2/app.py

The application will be available at http://127.0.0.1:5000/

Advanced Semantic Search Engine

Version 1 — Python Script + Whisper

cd Advanced-Semantic-Search-Engine

# Install dependencies
pip install -r requirements.txt

# Run the search engine
python version_1/AI_SEO.py

Version 2 — Flask + ChromaDB (Recommended)

cd Advanced-Semantic-Search-Engine

# Install dependencies
pip install flask langchain-chroma langchain-huggingface chromadb sentence-transformers

# Add movie subtitle files (.txt) to version_2/subtitles/ directory
# A Marvel Cinematic Universe dataset (23 films) is included

# Run the app
python version_2/app.py

The application will be available at http://localhost:5000/

OWASP AI Vulnerability Scanner

Setup & Run

cd owasp-ai-vulnerability-scanner

# Install dependencies
pip install flask requests

# Make sure Ollama is installed and running
ollama pull llama3.2:latest
ollama serve

# Run the app
python app.py

The application will be available at http://localhost:5000/ Health check endpoint: http://localhost:5000/health

AI Travel Planner

Version 1 — Streamlit + Gemini

cd Travel_planner

# Install dependencies
pip install streamlit langchain langchain-google-genai

# Set your Google API key
export GOOGLE_API_KEY="your-api-key"          # Linux/Mac
$env:GOOGLE_API_KEY="your-api-key"            # Windows PowerShell

# Run the app
streamlit run version_1/Travel_Planner.py

Version 2 — Flask + Gemini + Ollama

cd Travel_planner/version_2

# Install dependencies
pip install flask ollama google-generativeai python-dotenv requests

# Create .env file with your Gemini API key
echo GEMINI_API_KEY=your_key_here > .env

# Pull the Llama 3.2 model (requires Ollama running)
ollama pull llama3.2:latest

# Run the app
python app.py

The application will be available at http://localhost:5000/

Medical Receptionist Voice Agent

Setup & Run

cd Voice_agents

# Install dependencies
pip install flask flask-cors requests python-dotenv elevenlabs speechrecognition pyaudio

# Create .env file with API keys
echo GROQ_API_KEY=your_groq_key_here > .env
echo ELEVENLABS_API_KEY=your_elevenlabs_key_here >> .env

# (Optional) Create data.json with patient records — sample data is used by default

# Run the app
python main.py

The application will be available at http://localhost:5000/

Development Journey

Each project follows a dual-version architecture (where applicable) that documents a deliberate learning progression:

Phase	Version 1 (Assisted Build)	Version 2 (Self-Built)
Approach	Built with guidance from ChatGPT and tutorials	Developed independently from scratch
Frontend	Streamlit / CLI (rapid prototyping)	Custom HTML/CSS/JS (production-style)
LLM / ML	Google Gemini API (cloud), Whisper	Ollama + Llama 3.2 (local), HuggingFace, Groq
Architecture	Single-file, API-dependent	Modular Flask app with utils & templates
Key Learning	API calls, prompt structuring, quick prototyping	Full-stack development, backend routing, HTML templating, local LLM deployment

This dual-implementation approach demonstrates:

Different architectural choices for AI-powered applications
Progression from cloud APIs to locally-hosted models
Growth from assisted development to independent full-stack builds
Enhanced prompt engineering and error handling in self-built versions
Exploration of vector databases and semantic search pipelines
Application of AI to cybersecurity (OWASP vulnerability scanning)
Multi-model pipelines combining cloud + local LLMs (Travel Planner)
Voice-to-voice AI agents with TTS and speech recognition

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Fork the repository
Create your feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

Acknowledgements

LangChain — LLM orchestration framework
Streamlit — Rapid ML/AI app prototyping
Flask — Lightweight Python web framework
Google Gemini API — Cloud-hosted large language model
Ollama — Run open-source LLMs locally
Meta Llama 3.2 — Open-source language model
Groq — Fast AI inference API
ElevenLabs — AI text-to-speech platform
ChromaDB — Open-source vector database
Sentence Transformers — State-of-the-art text embeddings
HuggingFace — ML model hub and transformers library
OpenAI Whisper — Speech-to-text model
OWASP — Open Web Application Security Project
World Time API — Timezone and local time data
Open Exchange Rates — Currency exchange rate API

Name		Name	Last commit message	Last commit date
AI_DS_Tutor		AI_DS_Tutor
AI_code_reviewer		AI_code_reviewer
Advanced-Semantic-Search-Engine		Advanced-Semantic-Search-Engine
Travel_planner		Travel_planner
Voice_agents		Voice_agents
owasp-ai-vulnerability-scanner		owasp-ai-vulnerability-scanner
README.md		README.md

Latest commit

History

Folders and files

Repository files navigation

GenAI Playground

Table of Contents

Technologies

Projects

AI Data Science Tutor

Features

Version 1 — Streamlit + Gemini + LangChain

Version 2 — Flask + Ollama (Self-Built)

AI Code Reviewer

Features

Version 1 — Streamlit + Gemini API

Version 2 — Flask + Ollama (Self-Built)

Advanced Semantic Search Engine

Features

Version 1 — Python + Sentence Transformers + Whisper (AI-Assisted)

Version 2 — Flask + LangChain + ChromaDB (Self-Built)

OWASP AI Vulnerability Scanner

Features

AI Travel Planner

Features

Version 1 — Streamlit + Gemini + LangChain (AI-Assisted)

Version 2 — Flask + Gemini + Ollama (Self-Built)

Medical Receptionist Voice Agent

Features

Repository Structure

Getting Started

Prerequisites

Clone the Repository

AI Data Science Tutor

AI Code Reviewer

Advanced Semantic Search Engine

OWASP AI Vulnerability Scanner

AI Travel Planner

Medical Receptionist Voice Agent

Development Journey

Contributing

Acknowledgements

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages