Unable to permanently store files uploaded to RAG_API

[sk365]

What happened?

When uploading a file via the agents / filesearch api, upoladed file ends up in a temporary folder as

docker exec -it LibreChat sh
 /app $ find / -name "test3.txt"
/app/uploads/temp/67c81d9ee0ad8a11c560ac1a/test3.txt

File can then be used correctly during this session, but cannot be accesed after contaiiner restart (even if it appears in the list of uploaded files)

from my docker compose yml

> services:
>   api:
>     container_name: LibreChat
>     environment: 
>       - UPLOAD_PATH=/app/data/uploads 
>     volumes:
>       - ./uploads:/app/uploads/temp
>       - ./data/uploads:/app/data/uploads  # Host-Mount für API
> 
>   rag_api:
>     container_name: rag_api
>     image: ghcr.io/danny-avila/librechat-rag-api-dev-lite:latest
>     environment:
>       - DB_HOST=vectordb
>       - RAG_PORT=${RAG_PORT:-8000}
>       - UPLOAD_FOLDER=/app/data/uploads
>     env_file:
>       - ./.env
>     volumes:
>       - ./data/uploads:/app/data/uploads  
> 

This is not a permissions issue, error still occurs after

sudo chmod 777 ./data/uploads  

Version Information

Version Info:

 docker images | grep librechat
ghcr.io/danny-avila/librechat-dev                latest    56ca144f4ce7   5 days ago      876MB
ghcr.io/danny-avila/librechat-rag-api-dev-lite   latest    6550e7ddf180   10 days ago     1.3GB

Steps to Reproduce

1. Upload file to agent
2. Restart container
3. Try to retreive file content via Agent rag_api

What browsers are you seeing the problem on?

Firefox

Relevant log output

File upload is seen in 
docker logs rag_api
2025-03-17 17:12:55,836 - root - INFO - Request POST http://rag_api:8000/embed - 200
2025-03-17 17:24:16,187 - httpx - INFO - HTTP Request: POST https://api.openai.com/v1/embeddings "HTTP/1.1 200 OK"
2025-03-17 17:24:16,208 - root - INFO - Request POST http://rag_api:8000/embed - 200

no error entries in docker logs LibreChat

Screenshots

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[danny-avila]

Unable to permanently store files uploaded to RAG_API

This is not the expected design of the RAG API. Files are transformed to vectors then discarded. If you expect to re-use the same static files, look into OCR:

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

[sk365]

Thank you for your superquick reply, @danny-avila

OCR allows only to upload pdf, not txt or markdown files.
I would find it appropriate to have uploaded plain test (or markdown or json) files permanently usable for RAG.
Can I, or could you implement this?

Or did I miss a point about the RAG functionality?

[sk365]

PS: would this work using Mongo DB instead of Postgres (as stated here)
if so: How to implement?