What's your tech stack?
**Rozz’s tech stack**
Rozz is built as a modern, secure, and privacy‑conscious web‑component that powers AI‑driven search on any website. Its core stack comprises:
| Layer | Technology | Purpose |
|-------|------------|---------|
| **Front‑end component** | Lit (Google’s web‑components library) | Provides a lightweight, isolated web‑component that can be dropped into any site without affecting its DOM or performance. |
| **Embedding & semantic search** | Custom embeddings engine (likely based on open‑source LLMs such as GPT‑3/GPT‑4 or smaller models) | Converts website content into vector embeddings for semantic search and natural‑language query handling. |
| **Data extraction & structuring** | Proprietary crawler + Schema.org generation | Crawls the site’s HTML, PDFs, and other documents, then auto‑generates Schema.org QAPage markup and `llms.txt` files to expose structured data to AI agents. |
| **Security & privacy** | Hardened against XSS, prompt injection, referer leakage; operates only on public data | Ensures the component is safe to embed and does not leak sensitive information. |
| **Accessibility** | Semantic HTML5, ARIA attributes, WCAG 2.1 AA compliance | Full keyboard navigation, screen‑reader support, high contrast, and multilingual support. |
| **Dashboard & analytics** | Custom web dashboard (React/Next.js style) | Lets site owners monitor performance, refine responses, and analyze user behavior. |
| **Hosting & delivery** | Edge‑optimized CDN (e.g., Cloudflare Workers or Vercel Edge) | Delivers the component and processed embeddings with sub‑200 ms latency worldwide. |
**Sources**
- [Rozz Chatbot Security & Privacy](https://rozz.site/qna/rozz-chatbot-security-and-privacy.html) – details on Lit component, security safeguards, and public‑data approach.
- [Rozz Accessibility Features](https://rozz.site/qna/what-are-the-accessibility-features-of-rozz) – outlines semantic HTML, ARIA usage, and WCAG compliance.
- [Rozz Dashboard](https://rozz.site/qna/introducing-the-rozz-dashboard.html) – shows the analytics stack for monitoring and refining responses.
**Follow‑up**
Which part of the stack are you most interested in (e.g., embedding model choice, integration process, or security details)?