Nice! I wrote a script similar to this that triggers when I take a screenshot with Flameshot, OCR's the screenshot using an API and opens web page with the OCR'd text. I prefer to use a Japanese to Japanese dictionary when possible which I have as a browser extension (yomitan). I like that this approach doesn't require an external API.