Knowledge Base & RAG

What is RAG?

RAG stands for Retrieval-Augmented Generation. Instead of answering from memory, ARIA first retrieves relevant paper excerpts from a curated knowledge base, then uses those as context to generate your answer — with real citations.

How It Works

1. You ask a research question
2. ARIA converts your question to an embedding and searches the knowledge base using semantic similarity
3. The most relevant paper excerpts are retrieved with titles, authors, and PMC IDs
4. These excerpts are injected into Claude’s context alongside your question
5. Claude synthesizes an answer grounded in the retrieved papers with verifiable citations

STXBP1 Knowledge Base

17,000 curated PubMed Central papers (Jan 2000 – Jun 2025) plus 165 base editing analysis entries. Covers Munc18-1 protein function, seizure treatments, gene therapy research, variant studies, and therapeutic approaches.

SNAP25 Knowledge Base

2,043 curated PMC papers plus 683 OpenFold3 structural analysis reports and 53 expert-curated knowledge entries. Covers SNAREopathy research and synaptic vesicle dynamics.

Upload Your Own Documents

You can add medical records, genetic reports, or downloaded papers to a personal knowledge base. ARIA will index and include them in searches alongside the curated collections.