How to set local embedding model or use OpenAI key user provide to support files upload?

[piwawa]

I found that I can only write RAG_OPENAI_API_KEY in the .env file, but not automatically use the openai key inputed by the user like OPENAI_API_KEY=user_provided. This will result in an error message when the user uploads the document: Error processing file when i didn't set it. If I fill it in, my own key will be used for embedding. How can I automatically call the openai key entered by the user?

[piwawa]

?

[hayschan]

Does this documentation help you?

https://www.librechat.ai/docs/configuration/rag_api#default-configuration

[piwawa]

Does this documentation help you?

https://www.librechat.ai/docs/configuration/rag_api#default-configuration

I have read this POOR document before asking the question. This document is written ambiguously. What should I fill in? I tried set it like this:

RAG_OPENAI_API_KEY=user_provided
EMBEDDINGS_MODEL=text-embedding-3-small

But it throws an error directly and doesn't let me fill in the key!

Only setting my key in .env, it can work. Why can't just use the key what openai key user provided??