What is the best database for Retrieval-Augmented Generation (RAG)?

Community
1

I am developing an application where users upload documents and interact with them. Which database would be best suited for Retrieval-Augmented Generation (RAG) in my scenario

2

I recommend using Pinecone DB. You can apply for a free account that allows you to develop with a maximum of 100,000 records.

3

Hi @Saad1 , I would recommend to have a look at Chroma or one of the options available with LangChain. Here’s a link to the list.

4

In my use case user will be uploading so many pdf and docx files. So can I use FAISS? How to chose the best vector database? @vasyl @henryxiang

5

I came across another option that is Milvus. So can anyone please help me come to a decision of selecting the best vector database that can help me generate a good result?

6

Depending on your skills, I would recommend to set up a headless CMS (Directus) or SupaBase which both allow the user, file and API management connected to weaviate (either self-hosted or on cloud) as a vector DB service. Efficient, easy to work with and cheap.