scite Assistant uses a retrieval augmented generation (RAG) to find relevant research for your task, re-ranks sources, and then summarizes an answer using the pertinent documents.

By grounding sources from our database, we're able to 1) ensure we find relevant information and 2) create in-text references to real research we've indexed.

It's not just asking a question and generating an answer by guessing the next word but summarizing the information you need from research we found.

source: I'm the CTO of scite. Feel free to email me at ashish@scite.ai if you have more questions.