Learn how to connect your knowledge, configure your assistant, customize the widget, test your integration in staging, and deploy WebChat AI to your website.
Choose your path
Go from raw website content or uploaded files to a working assistant in 7 steps.
Start quickstart
Embed the lightweight SDK into HTML, React, Next.js, or single-page apps with origin controls.
View embed guide
Explore hybrid vector search, Reciprocal Rank Fusion, tenant isolation, and the REST API.
Explore RAG pipeline
Explore Documentation
High-level architecture from ingestion to grounded answers.
System Architecture
Multi-tenant backend with hybrid RAG and isolated client SDK
01. INGESTION
Websites & Documents
Playwright/HTTP crawler for sites, or multi-format parser for PDFs, DOCX, MD, and TXT files.
02. HYBRID RAG
Retrieval & Grounding
Cosine vector search + BM25-style lexical matching fused with Reciprocal Rank Fusion (RRF).
03. EMBED WIDGET
SDK & Origin Security
Self-contained 49 kB bundle rendering 11 theme presets and 10 fonts with domain allowlists.
From knowledge to answers
Watch Walkthrough
See how an assistant is created, customized, tested, and embedded in under two minutes.
Related documentation
Quickstart Tutorial
Step-by-step tutorial taking you from zero to a live embedded assistant.
Read guide
Knowledge Sources
Compare website crawling, uploaded files, and mixed ingestion modes.
Read guide
Widget Embed Guide
Learn script tag integration, framework mounting, and allowed domains.
Read guide
Customization Reference
Explore the 11 theme presets and 10 typography options.
Read guide
Ready to build?
Register a website and get a live assistant in minutes.
WebChat AI