That’s a really cool approach though! Maybe an updated version would be to have a headless browser shadowing your browsing and indexing the rendered HTML - there would still be some gaps though.