🎬 Vault Streaming

Vault Streaming is a lightweight, high-performance desktop application for discovering and managing cloud-streamed media. Originally built as part of the broader Vault Explorer ecosystem, it has been streamlined to focus entirely on cloud discovery, library tracking, and AI-powered media playback.

Features

• Streamlined Discovery: Scrapes TMDB/TVDB for rich cover art, backdrops, and cast info to help you find your next watch.
• Library Tracking: Keep track of the series and movies you are currently watching or have completed.
• Hardware AI Workflows: Leverage your local GPU for advanced, real-time video upscaling and audio processing capabilities.
• Live AI Subtitles: Enjoy real-time, multilingual AI transcription for your streams without the overhead of saving .srt files to your disk.
• Premium User Interface: Features a beautiful, minimalist UI with smooth transitions and customizable themes (Warm and Console modes) out of the box.

Development

This application is built with Electron and utilizes local FFmpeg and Python scripts for backend processing and AI workloads.

Tech Stack

• Frontend: Vanilla HTML/CSS/JS (No frameworks, pure performance).
• Backend: Electron (Node.js) IPC handling.
• AI Daemon: Python-based daemon managing NVIDIA NeMo (Parakeet-TDT) models for real-time transcription.
• Video Processing: FFmpeg integration for clipping, upscaling, and format conversion.

Quick Start

1. Ensure you have Node.js and npm installed.
2. Clone the repository and run:

npm install
npm start

🎥 AI Subtitles & Playback

Vault Streaming includes an embedded AI subtitle engine powered by the parakeet-tdt-0.6b-v3 model. When you start a stream, a Python daemon is spawned that processes audio via FFmpeg in real-time. Transcriptions are broadcasted directly to the Electron renderer via stdout, ensuring that no temporary .srt files clutter your disk while maintaining perfect sync with your video.

---

📋 FFmpeg Requirements

Vault Streaming uses FFmpeg for audio extraction and video manipulation.

Recommended Version: FFmpeg 6.0 or later Minimum Version: FFmpeg 4.4

Installation (Windows)

1. Download from: https://ffmpeg.org/download.html
2. Choose: Windows builds from gyan.dev
3. Extract to C:\ffmpeg and add C:\ffmpeg\bin to your system PATH.

Installation (macOS / Linux)

• macOS: brew install ffmpeg
• Ubuntu/Debian: sudo apt install ffmpeg -y

---

Theming & UI Architecture

The UI is built on a custom design system with two primary canonical shells:

• Console Mode: Deep purple/black aesthetic (#0b0813 background).
• Warm Mode: Light, approachable aesthetic (#F5F1E8 background).

All components are styled using CSS Custom Properties (--vault-* tokens). If you wish to create a custom theme, simply override these variables in your own stylesheet or through the in-app settings menu.

License

This project is licensed under the MIT License. See the LICENSE file for details.