Skip to main content
Sources are the content your bot learns from. This guide walks you through adding different types of sources and managing them effectively.

Adding a Source

1

Open Knowledge Base Settings

Go to your bot’s dashboard → select your server → AI FeaturesKnowledge Base
2

Click Add Source

Click the Add Source button in the Sources section.
3

Fill in the Details

Enter the required information based on your source type.
4

Save and Index

Click Save to add the source, then click Index to scan the content.

Source Types Explained

Single Page

Best for individual pages with focused content. Settings:
  • URL - The exact page URL (e.g., https://example.com/faq)
  • Title - A name for this source (e.g., “FAQ Page”)
  • Description - Optional notes about what this source contains
Example use cases:
  • Your FAQ page
  • A specific product guide
  • Terms of service
  • Pricing information
Single pages index fastest and are great for content that changes rarely.

Documentation Site

Best for multi-page documentation with linked content. Settings:
  • URL - The starting page URL (usually your docs homepage)
  • Title - A name for this source (e.g., “Product Documentation”)
  • Crawl Depth - How many pages to scan (1-5 levels deep)
How crawling works:
  1. The bot visits your starting URL
  2. It finds all links on that page
  3. It follows those links and scans each page
  4. This continues until it reaches your crawl depth limit or 100 pages
Example use cases:
  • Full product documentation
  • Help center articles
  • Wiki pages
  • Guide collections
Higher crawl depths take longer to index. Start with 1-2 and increase if needed.

Discord Channel

Best for using existing Discord content as a knowledge source. Settings:
  • Channel - Select a text channel from your server
  • Title - A name for this source
  • Include Bot Messages - Whether to learn from bot messages too
Example use cases:
  • Announcements channel (for product updates)
  • Rules channel
  • FAQ channel with pinned messages
  • Staff resources channel
The bot only indexes messages it can see. Make sure the bot has access to the channel.

Managing Sources

Viewing Source Status

Each source shows its current status:
StatusMeaning
PendingSource added but not yet indexed
IndexingCurrently scanning content
CompletedReady to use
FailedSomething went wrong (check the error message)

Re-indexing a Source

To update a source with new content:
  1. Find the source in your list
  2. Click the Refresh icon
  3. Wait for indexing to complete
Set up automatic updates (daily/weekly/monthly) if your content changes frequently.

Deleting a Source

To remove a source:
  1. Find the source in your list
  2. Click the Trash icon
  3. Confirm deletion
Deleting a source removes all its indexed content. The bot will no longer use it for suggestions.

Best Practices

Choose Quality Over Quantity

  • Do: Add your most helpful, accurate documentation
  • Don’t: Add everything and hope for the best
Quality sources with clear, well-written content work better than many sources with unclear information.

Use Descriptive Titles

Good titles help you manage sources later:
  • ✅ “Billing FAQ - Payments and Refunds”
  • ❌ “Page 1”

Set Appropriate Priorities

If you have multiple sources, set priorities:
  • 10 (highest) - Official documentation, FAQs
  • 5 (default) - General guides
  • 1 (lowest) - Community content, older guides
Higher priority sources are checked first when answering questions.

Use Categories for Accuracy

Limit sources to relevant ticket categories:
SourceCategories
Billing docsBilling, Payments
Technical guidesTechnical Support
Getting startedGeneral
This prevents the billing FAQ from answering technical questions.

Keep Content Updated

Outdated content leads to wrong answers:
  • Set automatic updates for frequently changing content
  • Manually refresh sources after major documentation changes
  • Delete sources for deprecated features

Troubleshooting

”Failed to index” Error

Common causes:
  1. Page not accessible - The URL might be wrong or the page requires login
  2. Page blocked crawling - Some sites block automated access
  3. Network issues - Temporary connectivity problems
Solutions:
  1. Verify the URL works in your browser
  2. Make sure the page is publicly accessible (no login required)
  3. Try again after a few minutes

”No content found” Error

Common causes:
  1. JavaScript-only content - Some pages load content with JavaScript
  2. Empty page - The page has no text content
  3. Login wall - Content requires authentication
Solutions:
  1. Check if the page content is visible without JavaScript
  2. Use a different page with actual text content
  3. Use a public-facing version of the page

Source Shows 0 Chunks

This means no usable content was found. Check:
  1. Does the page have text content?
  2. Is the content behind a login?
  3. Is the page structure unusual?
Try a different page or use a Discord channel source instead.

Need Help?

Get instant support via private DM