Apologies for the shameless self-promotion here but it was this very problem that I built puppeteer-heap-snapshot. It decouples the HTML from the scraper and in

Apologies for the shameless self-promotion here but it was this very problem that I built puppeteer-heap-snapshot. It decouples the HTML from the scraper and instead we inspect the booted app’s memory. Not near as performant but a lot more reliable. I wrote about it here: https://www.adriancooney.ie/blog/web-scraping-via-javascript-heap-snapshots https://www.adriancooney.ie/blog/web-scraping-via-javascript...