Can PDFSearchTool() read pdf's with images in them

Hi,

I frequently encounter failures in this tool particularly when there are images in the pdf file or pages that have been scanned. Is there a way to get around this?

Thx

I’m not sure how you can mitigate these issues, my first thinking is the tool doesn’t support images (scanned pdfs are images too).

If you don’t find a solution you might have to create a custom tool that uses a model with OCR capabilities to read such pdfs. Have you looked at https://llamahub.ai/ or LangChain tools? You might be able to find a better PDF tool that suits your needs.

Sorry I couldn’t give you a fix, I hope you find something :sweat_smile:

PDFSearchTool cannot read images because it does not have OCR capabilities.
You can try OCRTool or create your custom tool.