Skip to content

Frequently asked questions

General questions

What is RAGFab?

RAGFab is a knowledge assistant that allows you to ask questions in natural language about your internal document base. It uses RAG (Retrieval Augmented Generation) technology to search your documents and generate accurate answers with their sources.

What document formats are supported?

RAGFab supports the following formats:

  • PDF (including scanned PDFs via OCR)
  • DOCX (Word documents)
  • MD (Markdown)
  • TXT (plain text)

The answer does not match my question, what should I do?

  1. Check that you have selected the correct universe in the top bar.
  2. Rephrase your question more precisely.
  3. Enable Hybrid Search to combine semantic and keyword search.
  4. Rate the answer with thumbs down to help administrators improve the system.

The assistant cannot find information that exists in a document

  • The document may be in a different universe. Switch universes and ask again.
  • The document may not yet be ingested. Contact your administrator.
  • The relevant passage may be in a poorly extracted image or table. Report it to your administrator.

Account and login

I forgot my password

Contact your administrator to have your password reset.

How do I change my password?

Go to My profile and use the Change password button.

My session has expired

For security reasons, your session expires after a period of inactivity. Log in again with your credentials.

Conversations

I have reached the conversation limit

The number of conversations is limited (shown at the bottom of the sidebar, e.g. 50/50). To create new conversations:

  • Archive old conversations
  • Delete unnecessary conversations

How do I find an old conversation?

Use the search bar in the sidebar or check the Archives and Favorites tabs.

Administration

How do I add a new document?

See the Document management guide.

How do I give access to a new user?

See the User management guide.

What is re-ingestion?

Re-ingestion means reprocessing a document already in the base. This is useful after a document update or to improve chunking quality.