Skip to content
← Back to Guide AiServa Knowledge Base and RAG Lab

AiServa and AIKuari · AiServa and AIKuari: Local AI Guide

AiServa Knowledge Base and RAG Lab

VYROX Wiki · Updated 17 November 2026 · https://www.vyrox.com/wiki/aiserva-knowledge-base-and-rag-lab

The Knowledge Base in AiServa is where your organisation keeps the documents the AI answers from, and RAG Lab is a test page that shows how the AI searches them. They are used by staff who add documents and by managers who want answers they can trust. Both appear only when your administrator has given you access.

What It Does

  • Keeps documents in three spaces: Company, Private and Shared with Me.
  • Organises them into folders, with New Folder and Add Documents.
  • Shows each document as Ready, Failed or Not Indexed.
  • Searches a folder and shows the passages the AI would receive.
  • Lets managers check Health and see Documents Not Searchable in the Client Console.

Add Documents

  1. Open Knowledge Base and choose a space or folder.
  2. Press New Folder if you need one, and type a New Folder Name.
  3. Drop files on Drop Files Here or Browse, or press Add Documents.
  4. Wait until the status reads Ready. Only Ready documents can be searched.
  5. Use the Search tab and type a Question to check the right passages come back.

Note

You need the Contribute right to add documents and the Manage right to remove them. If a button is missing, ask a manager for access, or send an access request through Approvals.

Test a Search in RAG Lab

RAG Lab runs one question through every search stage and shows the result. It needs AI Agent access and Read on the knowledge base.

  1. Open RAG Lab.
  2. Pick a Knowledge Base and type a Question.
  3. Set Passages (Top 3, 5, 8 or 10), Minimum Similarity and Rerank.
  4. Press Run Search.
  5. Read the passages and the grounded answer with citations to your original documents.

Settings You Can Change

SettingMeaning
PassagesHow many passages go to the AI
Minimum SimilarityDrops passages that match too weakly
RerankOff, Knowledge Base Setting or AI Rerank with Scores

Other Tabs

  • Chunks and Vectors: pick a Document and press Show Chunks to see how it was split.
  • How It Works: explains the stages Extract, Chunk, Embed, Store, Search Two Ways, Fuse, Filter and Rerank, and Retrieve and Answer.

Frequently Asked Questions

Why Does the AI Not Find My Document?

Check its status. A document that is Failed or Not Indexed cannot be searched. Ask a manager to look at Health in the Client Console.

Does RAG Lab Change My Documents?

No. It only reads them and shows how the search works.

Where Do I Set Who Can Read a Folder?

Access and settings stay with managers in the Client Console.

Related Guides

  • AiServa Chatbots and Knowledge
  • AiServa Calendar, Approvals and Contacts
  • Ask AI in AiServa