Hey Guys,
I have a problem setting up something, let me explain.
So I have to upload pdf file into AI with large list of data (200+ items) to extract them and return it as a formatted JSON. Anthropic AI allows me only 8k output token so I can extract a maximum of 60 items.
What I do right now is to parse the document into text and ask AI to analyse it and return a JSON formatted response. To do this I nested one call into an API WF named “extraction”, and this is the one I’m calling so I get the JSON formatted.
I call it from another WF “save” : this one execute my extraction WF (so the nested call and return the formatted JSON), then execute a WF on a list ( to save the list of items extracted into my DB ), then schedule the WF save again.
I’m struggling to find the right condition for this WF to execute whenever all the items are extracted, I tried several things but I ended with only one extraction or infinite loop.
Here is a loom video for a more detail explanation :
Thank you very much in advance,
Hope a life saver shows up.