I am trying to use a custom gpt in chatgpt that takes an image-based pdf filled with invoices and uploads it to make.com using a webhook that connects to ILOVEPDF OCR Extraction and then once the file is created, send it back to the custom gpt. Is that possible. Each time i tried to upload the file to the webhook it was empty but testing on my terminal uploads it fine. I see it says I need to create a URL but I have no idea how to do that. Please can someone help me.
Did you create an action in your Custom GPT? I’m assuming yes, but just checking
If the file is created in ChatGPT, it’s put in a sandbox. You might need to turn on the data analyzer to be able to access that file. If you don’t you might not be able to do so.
At any rate, I found this in OpenAI’s documentation:
Thank you for your response and for that file. Everytime I try to upload the pdf it ends up truncating the basecode64 string and ends up uploading a corrupt file that can’t be used. The file is successfully uploaded with the properties, it is just that the basecode64 string is massively cut and well after that it can’t be used.
Can you share screenshots of your scenario and the configuration of the nodes?
Also, When you say that doing it with your terminal works, what do you mean? Do you mean taking that PDF and using something like Postman to upload it to your webhook? Or do you mean something else?