Can i create multiple index in standard plan?

Currently, I am using free plan and I am thinking to purchase ‘standard’.
Now, my question is, will pinecone me to create multiple indexes in that plan?
If do then I want to know if there are any limitations. let me know if pinecone set a particular number limit for the index in the standard plan.

Hey @abhishekpatel0044 you can create any number of indexes on paid plans. Just note that each index requires at least one pod, which will be billed.

There is a project setting called “Pod limits” which lets you limit the number of pods used in your project. This is useful for capping your cost and/or preventing unintentional scaling from you or your collaborators.

The default pod limit is 5. You can change this setting.

I want to build an AI chat service with the help of Langchain and Pinecone where people insert the website link and the AI will read the whole content of the site including every web page. After that visitors can ask the questions to AI chat so they can receive the appropriate answer from that content only.

I am storing all website data in one Pinecone index and for another website in I use another index. Basically, I am using one pinecone index for one website as of now. As code below.

pine = Pinecone.from_existing_index(web_1_index, embeddings)
chain = RetrievalQAWithSourcesChain.from_llm(llm=llm, retriever=pine.as_retriever())
res = chain({“question”: “what is xyz” }, return_only_outputs=True)

pine = Pinecone.from_existing_index(web_2_index, embeddings)
chain = RetrievalQAWithSourcesChain.from_llm(llm=llm, retriever=pine.as_retriever())
res = chain({“question”: “what is xyz” }, return_only_outputs=True)

And it will return the answer in the context of that website because that index has only that website data.

Now, what I am looking for is, I want to store two or more websites in one index; But, I want the reply in that particular website context (not two or many).

So the question is, Is there any way that I can store 2 or more website data in one index and get the answer in the form of one website context?

If yes, what is the drawback of this? OR do you have any suggestions like this are good practice or should I do only one index per site?
Also, I want to know how should I make this speed efficient and price efficient. (any idea for the improvement will be appreciated)

hey @greg and @abhishekpatel0044 , what i am trying to do is connect openai API with pinecone for history management. i have two or more users or many users for API , I have to store all their conversations in a database for context, and for faster retrieval and I also want to have different conversations for each user as we now have in chatgpt, so my question is should I create each index for each user ? just like we create a document for each user in MongoDB, and is it price and speed efficient ? please help me out

@john.daniel0499, that’s going to be very expensive to create an index per user, as Pinecone charges you a minimum price for each index. The right mechanism to logically separate data within an index is to use namespaces.

1 Like

This topic was automatically closed 24 hours after the last reply. New replies are no longer allowed.