🎥 YouTube Video Title Generator

Generate engaging, clickworthy YouTube video titles (with optional emojis and hashtags) powered by FastAPI, MongoDB Atlas Vector Search, and LLMs.

---

🚀 Features

• Generate optimized YouTube video titles from descriptions.
• Optionally add emojis and relevant hashtags.
• Ingest PDF data and create embeddings for vector search.
• Uses MongoDB Atlas vector search index with cosine similarity.
• FastAPI backend with auto-generated API docs.

---

📂 Project Structure

.
├── app
│ ├── config.py # Configuration and environment variables
│ ├── **init**.py
│ └── main.py # FastAPI entrypoint
├── create_vector_index.py # MongoDB Atlas vector index creation
├── generate_response.py # LLM response generation
├── get_embeddings.py # Embedding generation helper
├── get_query_results.py # Query MongoDB Atlas with embeddings
├── ingest_data.py # Ingest & chunk PDF data
├── requirements.txt # Python dependencies
└── README.md # Project documentation

---

⚙️ Installation

1. Clone the repository

git clone https://github.com/your-username/youtube-title-generator.git
cd youtube-title-generator

2. Create and activate a virtual environment

python3 -m venv venv
source venv/bin/activate   # macOS/Linux
venv\Scripts\activate      # Windows

3. Install dependencies

pip install -r requirements.txt

---

🔑 Configuration

Create a .env file in the root directory:

# Application Settings
APP_NAME="FastAPI App"
APP_VERSION="0.1.0"
DEBUG=True

# Server Settings
HOST="127.0.0.1"
PORT=8000

# Database Settings
ATLAS_CONNECTION_STRING="mongodb+srv://<username>:<password>@cluster0.xjslrno.mongodb.net/?retryWrites=true&w=majority&appName=Cluster0"

# API Settings
API_PREFIX="/api/v1"

# Secret keys
HUGGING_FACE_ACCESS_TOKEN="***"

---

Usage

1. Ingest Data

python3 ingest_data.py

2. Create Vector Index

python3 create_vector_index.py

3. Generate Response from Query

python3 generate_response.py

Or run the FastAPI app:

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

---

📡 API Endpoints

Method	Endpoint	Description
GET	/	Root health check
GET	/info	Show app info from env
GET	/ingest-data	Ingest & chunk PDF into embeddings
GET	/create-index	Create MongoDB Atlas vector index
POST	/prompt	Generate YouTube title & hashtags

Example /prompt Request

{
  "video_title": "How to Learn",
  "description": "You need to manage time, prioritize tasks, and take notes",
  "include_emojis": "yes",
  "video_type": ".mp4",
  "generate_hashtags": "yes"
}

Example Response

{
  "message": "Prompt generated successfully!",
  "output": [
    {
      "title": "📅 Master Time Management: Prioritize Tasks & Take Notes",
      "hashtags": ["#TimeManagement", "#Productivity", "#YouTubeTips"]
    },
    {
      "title": "📝 Learn the Secrets to Effective Note-Taking",
      "hashtags": ["#Notetaking", "#StudyTips", "#YouTubeTutorial"]
    },
    {
      "title": "🎯 Boost Your Productivity: Prioritize & Organize Your Tasks",
      "hashtags": ["#ProductivityHacks", "#TaskManagement", "#YouTubeLifeHacks"]
    }
  ]
}

📖 API Documentation

Interactive API documentation is available and auto-generated:

• Swagger UI: http://localhost:8000/docs
  Explore and test all endpoints with an intuitive interface.

• ReDoc: http://localhost:8000/redoc
  Browse comprehensive, well-structured API reference documentation.

---

🛠️ Development Notes

• Runs on FastAPI + Uvicorn.
• Requires MongoDB Atlas with vector index support.
• Uses HuggingFace LLMs for text generation.
• Extendable for custom embeddings and retrieval-augmented generation (RAG).

---

📜 License

MIT License. Feel free to use and modify.

---

👨‍💻 Author

Developed by Robiul Hossain