Dear all, online in then AI Studio I tested and deployment a format that I would like my pdf processing to look like. As of now. I was able to build the pipeline up to the mistral node.
I can’t find a way to request a specific format my ocr, as it can easily be done in AI Studio online, by giving the expected json format. I tried a code node, http request, Mistral but I run into errors the whole time..
Does anyone have an idea why this is so difficult to get running?
I have a feeling that it’s easier / more intuitive to do in vs, code.. Which is what I will try to do in the meantime.. I attached a picture and would be curious what you guys think. Thank you in advance.
Have a look at the example provided in this post and let me know if this helps. You can use the Mistral Vision APIs to extract data from documents using a prompt. You should be able to instruct it on how you want the data to be returned.
If you are able to provide an example document and the json structure you expect, I can try and build an example workflow.
What are all the file formats you want to support? Depending what they are, you can either build the mistral prompt before you send it or build different workflows to deal with different kind of files. For example if you have structured excel files, those are easier and cheaper to just consume using the excel nodes etc
