Hello everyone,<p>I am the developer of eazyrag.com, and here is the problem I am trying to solve.<p>When I was working on usegrasp.com (a search engine), I integrated the LLM answer engine, which is basically implementing a retrieval augmented generation (RAG) pipeline. First, I tried the most popular libraries available. I have to say I disliked the developer experience due to confusing APIs and complex abstractions. I wanted something like Algolia for retrieval augmented generation.<p>So, I built eazyrag.com, which is an easy-to-use API to implement RAG with your own data inside your apps or websites. You can simply index everything and query it. You don't even need a unique ID for the content you are indexing; just send the entire documents in a single API call, and we will handle chunking, splitting, embedding, and prompt with context formation, etc.<p>Here is a demo I built: I indexed all the Bun.js docs/guides pages on EazyRAG and performed RAG on them: <a href="https://eazyrag.com/bun" rel="nofollow noreferrer">https://eazyrag.com/bun</a>
Thanks for building this. For those who are not familiar with Retrieval Augmented Generation (RAG). Here are some resources<p><a href="https://leaptable.co/wiki/Retrieval-Augmented-Generation-(RAG)-for-Large-Language-Models-Explained-8c18bae88e8041c685f6c24e33a40a97/" rel="nofollow noreferrer">https://leaptable.co/wiki/Retrieval-Augmented-Generation-(RA...</a><p><a href="https://leaptable.co/wiki/Retrieval-Augmented-Generation-(RAG)-Architecture-8b0d22d397a6434cb37fb903d7ff8deb/" rel="nofollow noreferrer">https://leaptable.co/wiki/Retrieval-Augmented-Generation-(RA...</a><p><a href="https://leaptable.co/wiki/Effects-of-Chunk-Sizes-on-Retrieval-Augmented-Generation-(RAG)-Applications-8b728c36d005434dba39ad19be9b82cc/" rel="nofollow noreferrer">https://leaptable.co/wiki/Effects-of-Chunk-Sizes-on-Retrieva...</a>