Self host or cloud n8n. Http request the url over, use an http request node to scrape the site, use an ai agent node (Gemini 2.0 flash exp is free and quick) to look for the HubSpot script in the header, return structured JSON, webhook back to clay output table, use a lookup field to return the result back to the original table. Tons of tutorials on YouTube around it. Definitely worth the time investment and can save you a ton in the long run. Can be tweaked for other purposes besides Hubspot, while you do the scraping.