FAQ

General Questions

Q: What happens when I upload the same file twice?

A: When you upload a file with the same name as an existing vector, the system automatically overwrites the old data with the new version. The file remains the same, but all embeddings are updated.

Q: Can I edit my uploaded documents?

A: You cannot directly edit uploaded documents, but you can:

  • Delete and re-upload the corrected version

  • Reindex URLs to capture updated web content

Q: How long does vector processing take?

A: Processing time depends on:

  • File size and content length

  • Chunking configuration

  • System load

  • Typically 1-5 minutes for standard documents

Q: What’s the difference between Root and Groups?

A:

  • Root: Main directory containing all your vectors

  • Groups: Organizational collections that reference files in Root

  • Files in groups also appear in Root

  • Deleting from Root removes from all groups

Technical Questions

Q: Why is my search not finding relevant content?

A: Try these approaches:

  • Use different phrasing or synonyms

  • Check if content was fully processed (Completed status)

  • Verify chunking settings are appropriate for content type

  • Ensure you’re searching in the correct scope (Root vs. Groups)

  • You can find vector and rich content using the selected knowledge base with LLMs.

Q: Can I change chunking settings after upload?

A: No, chunking settings are applied during upload. To change settings:

  1. Note current settings

  2. Delete the existing vector

  3. Re-upload with new settings

Q: What happens to my vectors during system maintenance?

A:

  • Published chatbots continue using cached vector data

  • New queries may have limited functionality

  • System automatically recovers when maintenance completes

Q: How do I know if my vector data is being used?

A: Check integration points:

  • Flows: Look for {{vec-0.output}} variable usage

  • Chat: Responses will reference vector knowledge

  • Page: Content includes information from vectors

Integration Questions

Q: My chatbot isn’t using vector knowledge. What’s wrong?

A: Verify these settings:

  1. Vector sources are selected in Knowledge Source section

  2. Chatbot is published after configuration changes

  3. Vector processing is completed (not Processing status)

Q: Can I use vectors from different workspaces?

A: No, vector data is strictly isolated per workspace for security and data integrity.

Troubleshooting Content Issues

Q: Website crawling failed. What can I do?

A: Common solutions:

  • Verify the website URL is accessible

  • Check if the site has robots.txt restrictions

  • Try crawling specific pages instead of the entire sitemap

  • Some sites block automated crawling

Q: Vector processing stuck in “Processing” status.

A:

  • Wait 10-15 minutes for large files

  • Check if file content is supported

  • Contact support if status doesn’t change after 30 minutes

Last updated

Was this helpful?