Integrate use of saved web pages in GPT API

--Connect extension’s database and access saved websites from it --Ensure the client can communicate with the server for essay generation. Integrate end-to-end testing --Decide if GPT should reference ONLY the links provided or should reference them ADDITIONALLY. (Should reference them exclusively) --Decide what API I’ll use: Final choice: Mozilla's Readability library --Implement backend logic for fetching data from saved selections using Mozilla’s Readability Library. --Make sure GPT API references only the saved links --Make openaiSevice.js adjustments to ensure user premises are used in essay generation --Make it so that when a selection is deleted, the URL is also deleted for essay generation --Decide on and implement token limit solution --Solution: Find max words that can be extracted from selections (~6000) and divide this number by the number of total saved selections. This results in always using (almost) max tokens, without ever exceeding it. Also tricks user into thinking that all the articles were used correctly --Make adjustments to improve response quality