Plus you can probably use that until it fails (website changes) and then just re scrape it with llm request