Subgenie 🎥✨

Subgenie is an automated subtitle generation and embedding system for videos. It aims to make content more accessible for hearing-impaired and non-native speakers by providing an easy-to-use interface for generating and embedding subtitles in videos.

🔍 Problem Statement

• Videos without subtitles are inaccessible to hearing-impaired and non-native speakers.
• Manual transcription is time-consuming and resource-intensive.

🎯 Project Goal

Develop an automated system to:

• Generate subtitles from any language input
• Embed subtitles into videos
• Provide an easy interface for users to upload videos, automatically generate subtitles, and download the final video with embedded subtitles

🚀 Features

• 📤 Video upload and processing
• 🗣️ Automatic subtitle generation
• 📺 Subtitle embedding
• 🌐 Multi-language support
• ✂️ Video trimming and splitting
• 🎨 User-friendly design
• 🔄 Language translations
• ✏️ AI-generated subtitle editing

💻 Technologies Used

Frontend

• ⚛️ Electron.js: Enables desktop application features
• ⚛️ React: Used for creating the UI components
• 🎞️ FFMPEG: Used for video-to-audio conversion and embedding subtitles into videos

Backend

• 🐍 Python (or 🟩 Node.js): Handles MP3-to-SRT conversion and communicates with the frontend
• ⚡ FastAPI: For handling API endpoints to model
• 🗣️ Speech-to-Text Model: Converts audio into subtitle format (SRT)

🔄 System Overview

1. User uploads a video file
2. Audio is extracted and converted to MP3
3. MP3 is sent to the backend for speech-to-text conversion into SRT
4. The SRT file is sent back to the frontend
5. The SRT is embedded into the video, and the user can download the subtitled video

🏗️ System Architecture

📊 Data Flow Diagram

🎬 FFMPEG Architecture

🗣️ Whisper Architecture

⚡ Electron Architecture

🎯 Target Market

• 📹 Aspiring Vloggers
• 🌟 Social Media Stars
• 🚀 Startup Companies

🚀 Project Workflow

• Tech stack selection: Choosing the best tools for making native desktop apps
• UI Design Completion: User Interactive design in FIGMA
• Finding the Translational Models: Research on best language translational models
• Defining Features: Feature planning based on tech stack capabilities
• Development Started: Team split into frontend and backend development
• Completion of MVP: Development of all basic product features

🚧 Challenges Faced

• Burning subtitles to video using ffmpeg
• Limited access to premium translational models
• Deployment issues in various cloud providers