AniSubsAI

A desktop application for transcribing and translating audio files especially made for anime to generate subtitles, built with Python and PyQt6.

Overview

AniSubsAI provides a user-friendly interface to:

• Transcribe Audio & Video: Process both audio and video files, automatically extracting audio from videos.
• Multi-Engine Translation: Translate text using a variety of services, including Gemini, Google Translate, DeepL, and more.
• Flexible Subtitle Generation: Generate subtitles in multiple standard formats, including SRT, VTT, and ASS.
• Advanced Customization:
  • Select from a wide range of source and target languages.
  • Automatically detect the source language.
  • Choose your processing device (CPU/GPU) for transcription.
  • Easily switch between different faster-whisper models.

The application is designed with a modern, dark-themed UI and a modular architecture that allows for future expansion, such as adding more translation services.

Screenshot

Features

• Video and Audio Support: Transcribe directly from .mp3, .wav, .mp4, and .mkv files.
• High-Quality Transcription: Powered by faster-whisper for fast and accurate speech-to-text.
• Multi-Engine Translation: Choose between Gemini, Google Translate, DeepL, Microsoft Translator, and MyMemory.
• Automatic Language Detection: Let the application automatically detect the source language of your media.
• Multiple Export Formats: Save your subtitles as .srt, .vtt, .ass, or plain .txt files.
• Drag and Drop: Easily add files to the application by dragging them onto the window.
• Device Selection: Choose between CPU and GPU for transcription to balance speed and resource usage.
• Customizable Models: Easily switch between different faster-whisper models (e.g., small, medium, large) via the .env file.
• Modern UI: A sleek, responsive interface built with PyQt6.

Setup and Installation

1. Clone the repository:

   git clone <repository-url>
   cd AniSubsAI

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

4. Configure your environment:

   • Create a .env file in the root of the project.
   • Add your Gemini API key to the .env file if you plan to use the Gemini translator:

     GEMINI_API_KEY="YOUR_API_KEY_HERE"
     

   • (Optional) Customize the WHISPER_MODEL and GEMINI_MODEL in the .env file.

5. (Optional) GPU Support: For GPU-accelerated transcription, you will need to have the NVIDIA CUDA Toolkit installed and then install the following dependencies:

   pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

How to Run

Once the setup is complete, you can run the application with the following command:

python main.py

Project Structure

The project is organized into the following directories:

• core/: Shared utilities and configuration.
• gui/: All PyQt6 GUI components, stylesheets, and workers.
• models/: The directory where faster-whisper models are downloaded and stored.
• transcriber/: The faster-whisper transcription logic.
• translator/: The Gemini translation logic.
• main.py: The main entry point for the application.