Enhancement: Add Google Gemini Context Caching

[fret423]

What features would you like to see added?

Google has launched Context Caching recently to its Gemini products. This significantly reduces the cost of queries where there are long contexts. One of the ways that we use LibreChat internally is that we feed it our documentation (using the Files sidebar for RAG), and ask questions. Since the entire context of the documentation is fairly stable, it would be great to pay lower costs for it through caching.

More details

Context Caching is implemented by providing an expiry date, so this would need to be added somehow. I think it would be best by giving each file an option to cache through the File Manager modal.

Which components are impacted by your request?

General

Pictures

No response

Code of Conduct

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

[tfriedel]

also anthropic added the same feature yesterday

[danny-avila]

also anthropic added the same feature yesterday

Thanks, I am aware. Given that this is a totally different feature/provider, a new issue is in order: #3661