Quenvyra is a multi-agent RAG app that answers customer and product-review questions from your own store data. The AI backend and a customizable React front end ship together in one container: light with SQLite, or enterprise with PostgreSQL.
Stop jumping between the Shopify admin, review apps and policy docs.
Exact match by email or order number, plus natural-language search across orders, tickets and policies.
Aggregate reviews into answers: sizing trends, recurring defects and what customers actually say.
Specialised agents for orders, reviews and knowledge cooperate to ground every answer in your store data.
LLM synthesis turns retrieved context into a direct reply your support agents can send or edit.
Every row and vector carries the shop domain, and every query filters on it.
Built as a Shopify add-on that works with headless Hydrogen storefronts and webhooks.
One click drafts a reply grounded in the order, past tickets and your policies. Agents review and send.
Add, edit and delete policies, FAQs and product docs. Articles are chunked and re-embedded automatically.
Probe how the agent handles refund, off-topic and risky requests before customers see it.
A customizable React + TypeScript front end is bundled in the same container: an embeddable shopper widget (order lookup, ask a question, submit ticket) and a team admin console.
Tickets, Knowledge Base, Dashboard, Preferences and Guardrail Test in one place.
A fully customizable, shopper-facing front end built as a TypeScript React + Vite component. Drop it into your Shopify Hydrogen app and style it to your brand.
// app/routes/support.tsx (Hydrogen)
import { QuenvyraWidget } from '@quenvyra/widget'
import '@quenvyra/widget/style.css'
export default function Support() {
return (
<QuenvyraWidget
apiUrl="https://your-host:8000"
shopDomain="your-shop.myshopify.com"
tabs={['order', 'ask', 'ticket']}
theme={{ primary: '#4fbfb0', radius: 12 }}
/>
)
}Webhooks and knowledge in, grounded answers out.
Shopify webhooks ──▶ API ──▶ SQL store (orders, customers, tickets, reviews)
│
Knowledge articles ──▶ chunk ───┤
(policies, FAQs, docs) ├─▶ Vector store (orders · tickets · knowledge)
│
Agent query ──▶ /search/ask ────┴─▶ similarity search ──▶ LLM synthesis ──▶ AnswerExact lookups by order number or email.
Natural-language search across all sources.
Works with mock fallbacks until you add an API key.
Same app, same container image. Pick the storage that fits your traffic.
One container, zero infrastructure
Scale with PostgreSQL
Each license is valid for one Shopify shop.
Pull the image from GitHub Container Registry and run.
docker run -p 8000:8000 -v ./data:/app/data \ -e OPENAI_API_KEY=sk-... \ ghcr.io/quenvyra/support-rag:latest
docker run -p 8000:8000 \ -e DATABASE_URL=postgresql://user:pass@db:5432/quenvyra \ -e OPENAI_API_KEY=sk-... \ ghcr.io/quenvyra/support-rag:latest
API on localhost:8000, interactive docs at /docs, admin console at /admin, customer widget at /client.
<script src="https://your-host/widget.js" defer></script> <quenvyra-widget shop="your-shop.myshopify.com"></quenvyra-widget>
Deterministic fitment data decides compatibility; the agent explains why a part fits in plain language.
Fit feedback from reviews and returns becomes sizing guidance and proactive chat on hesitation.
Tell us about your store and we will help you pick the right edition.