Build a scraper using n8n and walk me through functionality

Hello,

I have built a flow that scrapes ebay and inputs details into an Airtable. It works by inputting a search URL and outputs the key fields/info on the page into Airtabe.

There are few things to improve - e.g. it doesn’t always get every result and I would also like the same scraping functionality replicated for 2 other sites.

I’m eager to learn myself so once you have done it, I’d need c. 1hrs of explanation on how it works.

Could you please send a PM if you are interested in the work and give me an idea of the charges/hourly rate and I can follow up with you directly.

Thank you.

@automatron while I am not able to assist in the 1:1 session, just wanted to share some hopefully useful infos for you:

When it comes to inconsistency between executions, I’ve found it’s often due to two things:

  • The website is using some sort of tech to prevent requests. If that’s the case, PhantomBuster is your best bet (and we have a node for it). I have not personally used it, but a few of our (super smart) growth hacker/ no-coder users have recommended it to me for these cases.
  • The website is using some sort of framework where the css/ html you’re seeing was “compiled” and so between builds will look a bit differently. No expert on this but from my experience usually lots of random alphanumerics in css classes points toward this. Gets a bit tricky here to find a unique selector, so brushing up on the long-tail of css selectors can be helpful (https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_Selectors), especially since these tools might have custom data attributes where otherwise there are no unique css classes.