SnapStudySensei is a tool to assist with capturing and extracting Japanese text, looking it up in a dictionary, and optionally recording flashcards into Anki.
This project is inspired by Game2Text but takes a different technical approach.
- Currently only supported on Linux. Porting it to macOS and Windows should
be doable by adding the ability to list windows (see
windows_list.py, patch welcome) - No configuration, only designed for my own needs so far
- A bit slow to start due to the OCR model initialization
Install Anki and its Anki-Connect plugin if you want to record flashcards. They are not required for the capture, OCR, dictionary, and text-to-speech features.
python -m venv venv
. venv/bin/activate
pip install -e .This is a non-intrusive standalone installation, which means the Qt libraries
are duplicated within the virtual env. This breaks at least the fcitx input
method setup on the system. Setting IBUS_USE_PORTAL=1 QT_IM_MODULE=ibus in the
environment can be used as a workaround (ibus doesn't need to be installed).
To use the flashcard features, start Anki with its Anki-Connect plugin before starting SnapStudySensei. If Anki-Connect is unavailable, SnapStudySensei starts without the Records panel and card-recording controls.
Enter the venv (. venv/bin/activate) if you used that installation method, and
run sss.
SnapStudySensei will automatically create a model, deck and flashcard templates in Anki. The deck is called SnapStudySensei and is located in the Japanese category.