Introduction

Understand what DGbot is, how it works, and what you can build with it in the next few minutes.

~5 minutes Free

What is DGbot?

DGbot is an AI chatbot platform that trains on your own content — websites, PDFs, FAQs, and documents — and deploys to your website as a chat widget. Visitors ask questions in natural language and DGbot answers using the information you've uploaded.

Unlike generic AI assistants, DGbot only knows what you teach it. This means answers stay on-topic, accurate, and grounded in your actual business content.


How it works

DGbot combines three systems working together:

Knowledge base — You upload your content (website pages, PDFs, FAQs). DGbot processes it, breaks it into chunks, and stores it in a vector database optimised for semantic search.

AI reasoning — When a visitor asks a question, DGbot searches your knowledge base for the most relevant chunks, then asks a large language model (LLM) to compose a natural-language answer. It only answers from what it finds — it does not guess or hallucinate from its training data.

Chat widget — A lightweight JavaScript widget installs on your website with one script tag. It connects to DGbot over WebSocket for real-time streaming responses.

Which AI model does DGbot use?
DGbot supports multiple models: OpenAI GPT-4o mini (fast, low cost), GPT-4o (most capable), Claude Haiku (efficient), and Claude Sonnet (premium). You choose the model per chatbot in Bot configuration → AI model. On the BYOK plan you bring your own API key.

What you can build

DGbot handles a wide range of use cases out of the box:

  • Customer support bot — answers product questions, pricing queries, and policy explanations 24/7 without human involvement
  • Lead generation bot — collects visitor name and email before or during a conversation, then sends the lead to your CRM via webhook
  • E-commerce bot — connects to Shopify to look up order status, search products, and check inventory in real time
  • Internal knowledge bot — trained on internal wikis, SOPs, and HR documents to answer staff questions
  • Developer assistant — trained on your API docs and SDK references for developer-facing support

Next steps

The fastest path to a working bot is the Quickstart guide — it takes about 15 minutes and requires no code.