This is a comprehensive Flask web application designed to process, stream, and automatically generate subtitles for video files. Users can either upload local videos or provide a URL for a video to be downloaded using yt-dlp. The application then generates subtitles using the powerful faster-whisper library, converts the video to HLS for efficient streaming, and provides a modern web player.

This project provides a user-friendly web interface for the core technology found in the faster-auto-subtitle project.

• Dual Input Methods:

  • Upload local video files directly through the browser.

  • Paste a URL from any yt-dlp supported site.

• Automatic Subtitle Generation:

  • Uses the faster-whisper library, a highly optimized implementation of OpenAI's Whisper model.

  • Option to upload a custom .srt file.

  • Advanced Control:

    • Choose from a wide range of Whisper models (from medium to large-v3-turbo).

    • Select the processing task: translate (to English) or transcribe (keep original language).

• Efficient Streaming:

  • Automatically converts videos to HTTP Live Streaming (HLS) format for adaptive bitrate streaming.

• Modern Web Interface:

  • Clean UI built with TailwindCSS and Alpine.js.

  • Automatic thumbnail generation for video previews.

  • List, play, and delete processed videos.

• Feature-Rich Player:

  • Modern Video.js player.

  • Tap/click on the sides of the player to seek forward or backward.

  • Subtitle support (.vtt format, converted on-the-fly).

Before you begin, ensure you have the following installed on your system:

1. Python 3.11 or newer:

   • This is required by the underlying faster-whisper library. Check your version with python --version.

2. FFmpeg: This is essential for video processing (thumbnailing, HLS conversion).

   • Windows: Download the binaries from the FFmpeg website and add the bin folder to your system's PATH.

   • macOS (using Homebrew):

     brew install ffmpeg
     
     

   • Linux (Debian/Ubuntu):

     sudo apt update && sudo apt install ffmpeg
     
     

3. (Optional, but Highly Recommended) NVIDIA GPU with CUDA 12:

   • To significantly speed up subtitle generation, faster-whisper requires CUDA 12 to run on a GPU.

   • You must install the NVIDIA CUDA Toolkit 12.x. Make sure the version is compatible with your GPU driver. For more details, see the faster-whisper installation guide.