SubPlz addresses the challenge of generating, synchronizing, and managing subtitle files for diverse media, with a particular focus on creating accurate audiobook subtitles akin to Kindle's Immersion Reading. It targets users seeking automated subtitle workflows, audiobook enthusiasts, and media library managers, offering a powerful toolset for enhancing media consumption and accessibility.

How It Works

The project employs a suite of advanced techniques, including AI-driven subtitle generation via faster-whisper and stable-ts, alongside sophisticated subtitle alignment algorithms like Alass and Needleman–Wunsch. Its core innovation lies in its flexible pipeline architecture, enabling users to generate multiple subtitle versions (e.g., AI-generated, aligned, original) and automate the entire subtitle processing workflow through a producer-consumer model.

Quick Start & Requirements

• Installation: Docker (recommended), from source, or via Google Colab.
• Prerequisites:
  • Docker with GPU support (Docker Desktop >= 4.44.2 required for WSL2).
  • NVIDIA driver with CUDA support (e.g., CUDA 12.x).
  • Python >= 3.11.2 (for source install).
  • ffmpeg (for source install).
  • A modern GPU with sufficient VRAM is highly recommended for performance.
• Links: [Colab Notebook](https://colab.research.google.com/drive/1_7z1_f50f33031217171