Using Kanzi AI tools

Kanzi provides AI tools that help you find information, understand Kanzi concepts and APIs, and get unblocked faster:

  • Ask Kanzi is the AI assistant built into the Documentation at https://docs.kanzi.com without any extra setup, just click the Ask Kanzi button at the top right of any documentation page to open it.

  • MCP servers connect your own AI assistant, such as Claude Desktop, VS Code, Claude Code, GitHub Copilot, and similar, to the same underlying Documentation and API references. Useful when you want to query Kanzi from inside your editor or chat tool.

Ask Kanzi

Ask Kanzi is grounded in the Documentation and API references, so it answers from verified content rather than from general training data. You can ask conceptual questions (“How do data sources work?”), how-to questions (“How do I bind a property to a data source value?”), API questions (“What is the signature of kanzi::Node::setProperty in Kanzi 4.1?”), or migration questions (“What changed in render passes between Kanzi 3.9.15 and 4.0.0?”).

Key things to know:

  • Page context. When you open Ask Kanzi from a documentation page, Ask Kanzi automatically uses that page as context, so you can ask follow-up questions about it without copy-pasting. Ask Kanzi shows the current page as a pill above the chat input. Click the × on the pill to remove the page from the context.

  • Selection context. Select any text on the page before opening Ask Kanzi, and that selection becomes additional context. This is useful when you want to ask about a specific code snippet, table cell, or paragraph.

  • Version. Ask Kanzi grounds its answer in the Kanzi version that matches the documentation page you opened it from.

  • Sources. Each answer lists the documentation pages it was based on. Click a source to open it in a new tab and verify the answer in context.

  • Find in chat. In the panel header, click the magnifier glass icon and type to find text in the current conversation. Use the up and down arrows (or Enter and Shift+Enter) to step through found terms. Press Esc or click the close button to clear the highlights.

  • New chat. In the panel header, click the pencil icon to clear the current conversation and start a new chat.

For details on how to use search across the same documentation, see Searching Documentation.

MCP servers

Kanzi MCP servers are built on the Model Context Protocol (MCP), an open standard for connecting AI assistants to external tools and data sources. After you connect a Kanzi MCP server to your AI assistant, the assistant can use the tools that the server provides to look up Kanzi information and answer your questions.

Available MCP servers:

  • Kanzi documentation MCP server connects your AI assistant to the Kanzi documentation. Use it to get answers about Kanzi concepts, features, workflows, and best practices.

  • Kanzi API MCP server connects your AI assistant to the Kanzi API reference. Use it to look up API signatures, check deprecation status, find include directives, and compare APIs between Kanzi versions.

Tip

You can set up the Kanzi Documentation MCP server in Claude Desktop and VS Code from two places on the Documentation website: from the AI assistant button next to any topic title, or from the Get Kanzi guidance in your IDE pill inside Ask Kanzi.

Tips for working with Kanzi AI tools

  • Be specific in your questions. For example, instead of asking “How do bindings work?”, ask “How do I create a binding in Kanzi Studio that changes the color of a node based on a data source value?”.

  • When asking about API use, mention the programming language and Kanzi version that you are using.

  • You can ask for verification of API signatures, check for deprecations, and suggest replacements for deprecated APIs.

  • When migrating between Kanzi versions, use the Kanzi API MCP server with your AI assistant to compare APIs between versions. This helps you identify what changed and estimate the migration effort.