This script records any tweets that appear in your Bookmarks page, and then saves them into a JSON file when requested.
You need to go to the Bookmarks page and then refresh the page for the script to start working. You will know that the script is working when the border around the bookmark list turns red.
The script will not download all bookmarks at once, only bookmarks that get fetched by the browser can be stored.
This means that you have to scroll down the list of bookmarks in order to capture them all. The scrolling speed can be quite fast, but it must be slow enough to ensure that the browser actually fetches the tweets. If you feel like you skipped some tweets, you can simply scroll back up to them and then continue on. (Protip: Use the middle mouse button to autoscroll while you make some coffee)
Press the F2 key to download all recorded bookmarks.
For testing, you can press F4 to print a list of all recorded tweets to the browser console.
The saved data contains the tweet URLs (which also include the tweet author), their text content, the date of creation, and the URLs of all embedded images. Videos/GIFs themselves are not included, but their existence in a tweet is. You can get the videos/gifs from these tweets using tools like yt-dlp by passing the tweet URL as an argument. Quoted tweets are not included
The script works purely by looking at the HTML of the page, it has no access to your account and it doesn't use Twitter's API.