Problems with loading text files

[mazierovictor]

I am facing a problem when loading files as text in the agents. It reports an error and when analyzing the docker log I realize that the error is in Mistral, as below.

2025-04-09 17:08:05 �[31merror�[39m: �[31m[getUserPluginAuthValue] No plugin auth OCR_BASEURL found for user 67df42a50dbd0630703ac8dc�[39m 2025-04-09 17:08:05 �[31merror�[39m: �[31m[getUserPluginAuthValue] No plugin auth OCR_API_KEY found for user 67df42a50dbd0630703ac8dc�[39m 2025-04-09 17:08:05 �[31merror�[39m: �[31mError uploading document to Mistral OCR API An error occurred while setting up the request: No plugin auth OCR_API_KEY found for user 67df42a50dbd0630703ac8dc�[39m 2025-04-09 17:08:05 �[31merror�[39m: �[31m[/files] Error processing file: Error uploading document to Mistral OCR API An error occurred while setting up the request: No plugin auth OCR_API_KEY found for user 67df42a50dbd0630703ac8dc�[39m

It turns out that I don't use Mistral AI, I only use OpenAI, including for RAG, as per the configuration below, .env:

#==================================================#

RAG

#==================================================#

More info: https://www.librechat.ai/docs/configuration/rag_api

RAG_API_URL=http://host.docker.internal:8000

RAG_OPENAI_BASEURL=

RAG_OPENAI_API_KEY=my_apy_key
EMBEDDINGS_PROVIDER=openai
EMBEDDINGS_MODEL=text-embedding-3-small
RAG_USE_FULL_CONTEXT=true
PDF_EXTRACT_IMAGES=true

[danny-avila]

uploading files as text concerns the OCR feature, not RAG

https://www.librechat.ai/docs/features/ocr

[mazierovictor]

I disabled the PDF_EXTRACT_IMAGES function, but when I try to load the file as text, I still get the same error.

Is there any way I can load the entire file for the agent to analyze?

I can't load the file just as a query.

[danny-avila]

Not with RAG

[mazierovictor]

Since not with RAG, what would be the option?

[danny-avila]

https://www.librechat.ai/docs/features/ocr

[mazierovictor]

@danny-avila

One more question...

If I configure Mistral OCR, in the Upload to File Search option, will it also be able to process images in PDF?