Create hourly parsed content records for knowledge base

Your hourly scraper returns raw page dumps that leave the chatbot with unsearchable, outdated content. Keep content current by storing parsed title, body, and excerpt in a table hourly.