rag-query

name	rag-query
description	Query a JS-PotatoRAG static vector database hosted on a static site (e.g. GitHub Pages). Use when the user points you at a JS-PotatoRAG site / index and asks to search their documents, retrieve context, or answer questions from their corpus. The site serves a prebuilt 3-bit polar-quantized index plus the js-vector-store library; you fetch the index, embed the query with the index's embedding model, and run a local top-k cosine search. Do NOT look for a query API — the search runs in your runtime.
version	1.0.0
license	MIT
homepage	https://github.com/MauricioPerera/JS-PotatoRAG

rag-query

This site is a read-only static vector database. There is no query endpoint. You retrieve context by downloading the index and the search library, then searching locally. The vectors never leave the static files; the cosine math runs in your process.

Inputs you need

• BASE — the site origin + path that serves the index, e.g. https://USER.github.io/JS-PotatoRAG (the dir that contains rag/ and js-vector-store.js). Take it from the user or from the homepage/origin.
• The user's natural-language .