Pls support Pinecone Retriever

[SeanJAmmirati]

We have checked your research assistant and it does a great job searching the web. However, we also would also like to have agents search across our Pinecone vector store.

From your blog post: https://blog.langchain.dev/exploring-uxs-besides-chat-with-research-assistant/ "It can equally easily be pointed to a vectorstore containing your relevant information."

But not sure when this will be done from your roadmap or if I'm just missing how to configure this. Any advice would be welcome.

[djuillard]

Hi, I have worked as well on this repo and the architecture, the sequencing are really smart.... and results are indeed very great -> impressive work. I also support @SeanJAmmirati request because in a market study for instance, it can make sense to add proprietary document into the global search in parallel to internet sources. Thanks

[SomethingGoodLLC]

How would PineCone help?

On Apr 4, 2024, at 05:17, djuillard @.***> wrote:



Hi, I have worked as well on this repo and the architecture, the sequencing are really smart.... and results are indeed very great -> impressive work. I also support @SeanJAmmiratihttps://github.com/SeanJAmmirati request because in a market study for instance, it can make sense to add proprietary document into the global search in parallel to internet sources. Thanks

— Reply to this email directly, view it on GitHubhttps://github.com/assafelovic/gpt-researcher/issues/379#issuecomment-2036631188, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AX7AYIUQX5EENBPBK2LZUBTY3ULD7AVCNFSM6AAAAABEJE4V3GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMZWGYZTCMJYHA. You are receiving this because you are subscribed to this thread.Message ID: @.***>

[SeanJAmmirati]

Because, those proprietary documentation (or even just specialized knowledge that may not be confidential) could be made accessible via searches against pinecone.

How would PineCone help? On Apr 4, 2024, at 05:17, djuillard @.> wrote:  Hi, I have worked as well on this repo and the architecture, the sequencing are really smart.... and results are indeed very great -> impressive work. I also support @SeanJAmmiratihttps://github.com/SeanJAmmirati request because in a market study for instance, it can make sense to add proprietary document into the global search in parallel to internet sources. Thanks — Reply to this email directly, view it on GitHub<#379 (comment)>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AX7AYIUQX5EENBPBK2LZUBTY3ULD7AVCNFSM6AAAAABEJE4V3GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMZWGYZTCMJYHA. You are receiving this because you are subscribed to this thread.Message ID: @.>

[SeanJAmmirati]

Does my explanation make sense? Any update on this?

[assafelovic]

Hey, although this is not my personal roadmap, happy to see if anyone from the community would be interested in this integration. Should be quite simple by adding another provider here: https://github.com/assafelovic/gpt-researcher/blob/master/gpt_researcher/memory/embeddings.py