Comment by _boffin_

Anyone know of any research on generating HTML differentials against updated webpages and with automatic healing of wrappers / selectors or research on using LLMs with webscraping and how to reduce token usage while retaining context?