Extending Possibility to Opt-in for Using Prompt Compression or using as a Tool.

[jatinchawda1503]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

Hello Everyone,

PromptCompression using LongLLMLingua or AutoCompresser refer this blog helps to compress the tokens by 6x-7x.

Using langchain, I have seen that the token consumption is steadily increasing especially when we use different types of Memory Types also mentioned in the pinecone blogs.

It would be great to introduce in the langchain core features to either opt in for the prompt compression or to use as an external function compatible with the current module.

This will help to optimize the cost for folks who are using OpenAI to build apps using LangChain.

Motivation

Enabling prompt Compression will shortern the original prompt length while keeping all the necessary information which helps to reduce cost of tokens while using 3rd party APIs.

Proposal (If applicable)

No response