Contents Menu Expand Light mode Dark mode Auto light/dark mode Auto light/dark, in light mode Auto light/dark, in dark mode Skip to content
Kanzi AI Assistant 1.5.0 documentation
Kanzi AI Assistant

Getting started

  • Kanzi AI Assistant 1.5.0
  • The chat window
  • Settings

Features

  • Viewport vision
  • Importing from Blender
  • Web search
  • MCP server

Reference

  • Release notes
  • License information
    • KANZI ONE SOFTWARE LICENSE TERMS
    • Third-party libraries used in Kanzi AI Assistant
Back to top

Web search¶

Web search lets Kanzi AI Assistant look up current information on the live web — recent release notes, library version changes, current best practices — for questions that the curated knowledge base does not cover. Kanzi AI Assistant prefers the knowledge base for Kanzi-domain questions and falls back to web search for the long tail.

Web search is off by default and requires an API key from a search provider. When it is off or unconfigured, Kanzi AI Assistant does not use it.

Enabling web search¶

  1. In Kanzi Studio, select AI Assistant > Settings.

  2. Select the Web Search category.

  3. Turn on Enable web search.

  4. In Search API key, paste your search-provider API key. By default Kanzi AI Assistant uses the Tavily search API, so a Tavily key works out of the box. Get a key from your provider’s portal.

  5. Optionally, set Search API URL to a different endpoint. Leave it blank to use the default Tavily endpoint.

  6. Click Save.

Kanzi AI Assistant stores your search API key locally in %APPDATA%, encrypted with Windows DPAPI under your user account, exactly like your LLM provider key.

Using web search¶

Once enabled, Kanzi AI Assistant calls web search on its own when a request needs up-to-date, off-corpus information. You do not call it explicitly — ask your question, and Kanzi AI Assistant decides whether to consult the knowledge base, the web, or both.

Next
MCP server
Previous
Importing from Blender
Copyright © 2026, Rightware Oy
Made with Furo
On this page
  • Web search
    • Enabling web search
    • Using web search