Using LangChain4j to analyze PDF documents

[giscus[bot]]

Using LangChain4j to analyze PDF documents

Quarkus: Supersonic Subatomic Java

https://quarkus.io/blog/using-langchain4j-to-analyze-pdf-documents/

[bulldog98]

Does anyone know if there is a local ai model, in case you have a customer wo does not want his or her data sent to a cloud provider.

[geoand]

AFAIK, very few models handle PDF documents

[syr]

@geoand any specific models you could recommend?

[syr]

@geoand any specific models you could recommend?

[geoand]

I've only used gemini-1.5-flash

[syr]

ok, me too :) with 2.0-flash btw.

[felipenova]

I have this error
java.lang.IllegalArgumentException: Unknown content type: PdfFileContent { pdfFile = PdfFile { url = null, base64Data = "JVBERi0xLjcNJeLjz9M
I cut the base64Data here because is big
I create the same example and i use deepseek but the error is before call LLM

[geoand]

PDF analysis only works with Gemini currently