🤖 BiQ ChatBot Demo

This is a RAG (Retrieval Augmented Generation) chatbot for the BiQ Group website. It scrapes websites, creates embeddings, and answers questions using AI about BiQ.

How It Works

  1. Scraping - Fetches content from a website and converts HTML to Markdown
  2. Chunking - Splits content into manageable chunks with hash-based change detection
  3. Embedding - Creates vector embeddings using Gemini for semantic search
  4. Indexing - Stores chunks and embeddings in Elasticsearch for fast vector search
  5. Chat - When you ask a question, it finds relevant chunks and sends them to the LLM

Technology Stack

ASP.NET Core 10 MSSQL Elasticsearch Semantic Kernel Gemini Embeddings PromptEnhancer RAG Pattern

Available Endpoints

/search

Semantic search API endpoint. Use ?q=your query to search across indexed content.

GET /chat/search?q=what is biq

/clear

Clears your chat session history and starts a fresh conversation.

GET /clear

/admin

Admin panel for managing scraped content and Elasticsearch index. Requires password authentication.

GET /admin