Hey @dot, welcome to the community!
I am sorry to hear you’re having trouble here. I am not super familiar with Grist, but I think we can nevertheless get this to work.
Assuming you are storing the URLs of all feed items in Grist I think you can avail of the Merge’s node Remove Key Matches mode. This allows you to only look at items not (yet) present in your Grist table and should help you avoid duplicates.
A workflow could look like so:
I’ve tested this and it seems to work as expected. Only new items pass the Merge node and only there are no duplicates:
Does this help?