yomisub is a desktop application for automatically generating subtitles for video files using AI-powered transcription with OpenAI Whisper. It is designed to assist with language learning by enabling users to watch foreign television shows and movies with native-language subtitles, even when official subtitles are unavailable.

The application includes:

• A GTK-based graphical user interface
• File picker for selecting video files
• Audio track selection for videos with multiple language tracks
• Whisper model selection (e.g., tiny, base, medium, large-v2)
• Subtitle generation in .srt format, saved alongside the video file

This tool was created to support language immersion and comprehension, particularly for learners of Japanese.

• Python 3.11

Tested on Debian with the following installation process:

sudo apt update


# install system dependencies

sudo apt install python3.11-venv ffmpeg python3-gi python3-gi-cairo gir1.2-gtk-3.0 libgirepository1.0-dev gir1.2-gtk-3.0

# add system site package to venv
python3 -m venv .venv --system-site-packages

# reload venv
source .venv/bin/activate

# test that gi can now be imported while in the venv
python -c "import gi"

# install pip dependencies
source .venv/bin/activate
pip install -r requirements.txt

python3 yomisub.py

Choose a model which will fit within the VRAM space of your GPU. Smaller models are faster but less accurate.

Whisper will download the chosen model to ~/.cache/whisper

This project is open source under the MIT License. See LICENSE.txt for full details.

© 2025 Sean Esopenko