When you execute recognize() function (ex: recognize(image, 'eng')), the language model to download is determined by the 2nd argument of recognize(). (eng in the example)
Tesseract.js will first check if *.traineddata already exists. (browser: IndexedDB, Node.js: fs, in the folder you execute the command) If the *.traineddata doesn't exist, it will fetch *.traineddata.gz from tessdata, ungzip and store in IndexedDB or fs, you can delete it manually and it will download again for you.