Project: UNIVERSE DOWNLOADER - Full-Stack Web Development | Pool Deza
Details, tech stack, key features, and impact of UNIVERSE DOWNLOADER
Professional media download platform with AI support and scalable architecture.
Universe Downloader is a full-stack web application that allows users to download videos and audio from multiple platforms (YouTube, TikTok, Twitch, Kick, Facebook) in maximum quality.
I developed a scalable solution with an intuitive interface, clean code layered architecture, and background execution with real-time monitoring.
As the System Architect, I was responsible for the complete architecture, integrating external services such as OpenAI Whisper for transcription and process optimization with FFmpeg.
- RoleSystem Architect
- Year2025
- CorePython 3.11, FastAPI
- AI ModelOpenAI Whisper
Support for YouTube (4K), TikTok, Twitch, Kick, and Facebook. MP4 formats (up to 2160p) and high-fidelity MP3.
Integration with OpenAI Whisper to convert downloaded audio to text automatically.
Monitoring of speed, ETA, and progress percentage using polling and websockets.
Asynchronous processing with FFmpeg to ensure universal video compatibility.
Implementation of threading + progress hooks to keep the UI responsive during heavy downloads without freezing the application.
Integrated a fallback mechanism for TikTok (using pybalt CLI if yt-dlp fails), ensuring high availability.
Executable compilation with PyInstaller including the Python runtime and dependencies for a 'Click & Run' experience.