hey everyone, i'm running a Claygent to get some data from the websites that are often dynamically rendered (JS) - mainly counting number of listings/cards or actually finding the explicit mention on the page. I've tried multiple prompts and models, those worked quite well with simple pages. But often fail with dynamic pages. I'm wondering if thats the prompt/model or overall Claygent limitation. Has anyone dealt with similar problem and do you have any tips? Really appreciate the help here!
One creative solution I can think of (outside of clay) is to vibecode a puppeteer/playwright script; then feed the image to claude API for analysis into a .csv to be uploaded back to clay. Takes time so Iβd do this only if itβs mission critical. There might also be better ways I canβt think of right now
Nice! Glad to hear:)
.png)