Add & maintain knowledge

Crawl your website

1. Open your chatbot’s Knowledge tab.
2. In Crawl your website, enter the public page where you want the import to begin.
3. Choose the page count offered by the screen and choose Crawl.
4. Wait for the result, then inspect Sources and the reported number of imported pages.
5. Test a question about that content in Install & test.

Good to know. A crawl currently accepts up to 15 pages per run on all plans; the initial selection is 8. It follows links on the starting domain. A lower actual page count can mean fewer reachable pages, blocked pages, or unreadable content—not a complete import of every page on your website.

Upload files

1. Use Upload files to choose PDF, TXT, Markdown (.md), CSV or HTML documents.
2. Wait until the files appear in Sources.
3. Test questions that depend on the uploaded information.

Good to know. Up to 50,000 extracted characters per file are retained. Image-only PDFs may have no readable text; use a text-based PDF or paste the content. Avoid uploading private customer data or secrets.

Paste text or Q&A

1. In Add text or Q&A, give the source a useful title such as Returns policy.
2. Paste clear business information, or use Q: and A: pairs.
3. Choose Add and check the source list.

Q: How do I contact your team?
A: Use the contact form on our website. Include your order number if your question is about an order.

Keep answers accurate

When your business information changes, remove outdated sources and add the updated text or files. Re-crawl the relevant website pages when needed; importing is not a scheduled automatic refresh.

Review the source list after repeated imports. Do not assume a new import replaces every earlier source. Remove old or conflicting versions, then test again.

Published FAQs are separate from knowledge sources. If an approved answer becomes outdated, edit or unpublish it in the Approval queue too.