Skip to main content

Contents & Improving AI Chatbot - LiveChatAI

Add Website Content to Your Chatbot

Import your website or help-center pages to give LiveChatAI a reliable, up-to-date knowledge base.

The crawler scans public URLs, breaks the text into context-aware chunks (AI Boost is always on), and trains your bot automatically—no manual copy-and-paste required.

Plan and Character Limits

Plan

Data-source quota

Basic

5 million characters

Pro, Advanced, Expert

10 million characters

Step-by-Step Guide

1. Open Data Sources → Add new → Website

The Website Source wizard appears.

2. Choose a Crawl Type

0:00
/

Option

When to use

All Pages

Crawl an entire subdomain (for example, help.example.com)

Single Page

Import one article or landing page

Sitemap URL

Crawl the URLs listed in your XML sitemap

3. Enter the URL and select Crawl pages

  • LiveChatAI processes up to 2,500 pages per crawl.
  • Crawling runs in the background, so you can continue working elsewhere in the dashboard.

4. Monitor Progress

The status column moves through three stages:

Crawling  →  Training  →  Trained

AI Boost is applied automatically to every crawled page.

5. Test Your Bot

Open Preview and ask questions that should be answered by the newly added pages. Confirm responses are correct.

Best Practices

Tip

Reason

Crawl support docs first

Resolves the most common unanswered queries

Use Exclude Paths

Skip blog posts, footers, or other low-value sections

Re-crawl after major updates

Keeps answers current; you can also enable Weekly Site Sync

Watch character usage

Large sites plus AI Boost consume your quota quickly

FAQ

How long does a 2 500-page crawl take?

 Typically five to twenty minutes, depending on page size.

Can I delete individual pages later?

Yes. Open the Website source, select the pages, choose Delete, then retrain the bot.

Does AI Boost cost extra?

 No. It is built-in and always on for Website sources.

Why did a page fail to import?

 The page must be publicly accessible, under roughly 2 MB of HTML, and not blocked by robots.txt.

Need assistance? Email [email protected] with the URL you’re trying to crawl.