Comic-AI (🌐Live Demo)

Turn any YouTube video or uploaded video into a comic strip using AI! 🎨 Transform your favorite moments into shareable comic art with just a few clicks.

---

🌟 Features

• Convert Videos to Comics: Turn YouTube videos or uploaded videos into 4-panel comic strips
• AI-Powered: Uses advanced AI models (OpenAI + Google Imagen) to generate high-quality comics
• Natural Language Prompts: Describe your comic in plain English
• Dual AI Engine: Primary and fallback AI models for maximum reliability
• Professional Quality: High-resolution 1024×1024 comic strips
• Easy to Use: Simple web interface with Streamlit
• Free & Open Source: No cost for setup and deployment

🎥 Demo

Input Video

Generated Comic

🔧 How It Works

1. Video Analysis: Uses Gemini AI to analyze video content and context
2. Prompt Enhancement: Transforms user descriptions into detailed AI prompts
3. Comic Generation: Generates comics using OpenAI's gpt-image-1 or Google's Imagen
4. Fallback System: Automatic fallback to secondary AI if primary fails
5. Web Interface: Streamlit-based UI for easy interaction

---

🚀 Getting Started

1️⃣ Prerequisites

• Python 3.8+
• YouTube video URL or video file
• API keys (see below)

2️⃣ Installation

Clone the repository

git clone https://github.com/SannidhyaDas/Comic-AI.git
cd Comic-AI

Create virtual environment (recommended)

python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate

Install dependencies

pip install -r requirements.txt

3️⃣ Configuration

1. Create a .env file in the project root:

GEMINI_API_KEY=your_gemini_api_key_here
OPENAI_API_KEY=your_openai_api_key_here

2. Get your API keys:

• Gemini API Key: Google AI Studio
• OpenAI API Key: OpenAI Platform

4️⃣ Running the App

Run the Streamlit app

streamlit run app2.py

Open your browser to http://localhost:8501

🛠️ Technical Architecture

---
config:
  look: neo
  layout: fixed
---
flowchart LR
    A["User Input"] --> B("YouTube Video URL + Comic Description")
    B --> C{"Video Analysis"}
    C -- Gemini API --> D("Prompt Enhancement")
    D --> E{"Comic Generation"}
    E -- "Primary: OpenAI gpt-image-1" --> F["Comic Strip"]
    E -- Fallback: Google Imagen --> F
    F --> G["Streamlit Web Interface"]
    G --> H["User Output"]

Loading

---

📦 Project Structure

Comic-AI/
│
├── assets/                        # Streamlit web interface
│   ├── appInterface_1.png            # Yt video & prompt example 
│   ├── appInterface_2.png            # generated comic example 
│   └── generated_img1.png            # comic example 
│
├── v1/             # Gemini free image gen model
│   ├── requirements.txt            # Python dependencies
│   ├── main.py             # Core logic and LLM pipelines  
│   ├── example1.ipynb      # example with functions of main.py 
│   └── app.py              # Streamlit user interface
│
├── v2/              # paid models(OpenAI + Gemini Imagen) 
│   ├── requirements.txt            # Python dependencies
│   ├── main2.py             # Core logic and LLM pipelines
│   ├── example2.ipynb      # example with functions of main2.py 
│   └── app2.py              # Streamlit user interface
│
└── README.md                   # Project documentation

📌 Version Information

This repository contains two versions of the application:

v1: Free Tier Implementation

• Uses gemini-2.0-flash-preview-image-generation (free image generation model)
• Limitations:
  • Text rendering quality is inconsistent
  • Image quality can be suboptimal
  • Text in speech bubbles may appear blurry or garbled

v2: Premium Implementation

• Implements a fallback strategy using paid models:
  • Primary: OpenAI's gpt-image-1 (superior text rendering)
  • Fallback: Google's imagen-4.0-generate-001 (high-quality image generation)
• Benefits:
  • Significantly improved text clarity
  • Higher overall image quality
  • More reliable comic generation

The deployed version uses v1 due to the payment requirements of the premium models. However, the v2 code is fully functional - users only need to provide their own API keys and have active billing with the respective services to use the premium models.

---

📬 Contact

Sannidhya Das - @LinkedIn - dassannidhya003@gmail.com

Project Link: https://github.com/SannidhyaDas/Comic-AI