I'm working on a Python script and I want to use Bedrock Claude 3.5 to extract information directly from PDF files. I've seen that in the playground, I can upload a PDF and specify what I need extracted, which works great. But I'm curious if it's possible to do the same thing from my script. Can anyone guide me on how to pass the whole PDF file to Claude 3.5 directly? Thanks for your help!
2 Answers
You might want to try this GitHub project: https://github.com/docling-project/docling. It could have the functionality you're looking for!
Check out this step-by-step guide on using AWS Bedrock and Claude 3.5 with Python: https://medium.com/@enriquecano12/getting-started-with-aws-bedrock-and-claude-3-5-sonnet-v2-a-step-by-step-guide-in-python-9929ba9b3775. It might give you some insights into your needs. Good luck!
Thanks! I actually have a script that works, but I’ve been converting my PDF to a string to pass it. I'm looking to see if I can pass the actual PDF file instead.