By scraping the h1 and h2 heading of like the top 10 serp one can do entity analysis on that too. Would be nice if you could select the number of serp results to be scraped. I would use it by scraping the h1 and h2 headers of the 10 first serp pages and drop all those headers in the entity analysis tool. Gives a much faster insight into major entities.

Maybe the scaleserp api is interesting for this use? https://trajectdata.com/serp/scale-serp-api/data-extraction/