Using PDFs in the chat hardly works

Hello everyone, I really like using hugging chat in my web browser because it allows me to easily integrate different language models and create my own assistants.

However, what I have noticed is that Hugging Chat has recently been having problems reading PDF files. Even if I use tools such as “Document”, “pdf-parser” etc. in an assistant, I always get an error message saying that the file could not be read. In the past, at least after a few attempts (model “CohereforAI”) it worked and the result made sense.

The problem exists both with PDFs that contain handwritten texts and with normal PDFs with machine-written texts.

Background:
I have created a wizard for students to get feedback on their essays in German. I have told the students that they can take a picture of their text and save it as a pdf and then upload it to the wizard.

Do you have any tips on how I can solve this problem?
Is there a “tool” that is particularly suitable for this?
Can Huggingchat also process images?

1 Like

I think the tools with camera icons are tools for inputting and handling images.
For now, I think it’s probably a bug, so I think it would help the developers if you reported it at one of the following.

I found this.