Prerequisites
- A knowledge base created in Swiftsell (go to Knowledge Base → New if you haven’t created one yet)
Add a domain (crawl)
Use a domain crawl when you want to index your entire website or a large section of it.- Open your knowledge base and click Add Source.
- Select Domain.
- Enter the root URL of the domain (e.g.,
https://www.yourcompany.com). - Click Add.
- The crawler respects
robots.txt. Pages blocked byrobots.txtwill not be indexed. - Use a subdomain or path-restricted URL to limit scope (e.g.,
https://help.yourcompany.comto only crawl your help centre). - Login-protected pages cannot be crawled.
Add a single URL
Use this to index a specific page — useful for important pages that you want to ensure are indexed regardless of crawl coverage.- Open your knowledge base and click Add Source.
- Select URL.
- Enter the full URL of the page (e.g.,
https://www.yourcompany.com/shipping-policy). - Click Add.
Upload a file
Use this to index documents that aren’t on the web — internal FAQs, product manuals, policy documents. Supported formats: PDF, DOCX, TXT- Open your knowledge base and click Add Source.
- Select File.
- Click Choose File and select your document.
- Click Upload.
Maximum file size is 10 MB. If your document is larger, split it into multiple smaller files before uploading.
Checking training status
After adding a source, it appears in the source list with a Pending status. Swiftsell processes the content asynchronously — refresh the page to see updates.| Status | What to do |
|---|---|
| Pending | Wait — processing is in progress |
| Completed | Source is indexed and ready to use |
| Failed | Check the URL is publicly accessible or the file format is supported, then try again |
After adding sources
Once at least one source showsCompleted:
- Attach the knowledge base to a Voice Agent — see Attach a Knowledge Base
- Use it in an AI Agent block in your automation — see AI Agent block