The Kajabi MCP is secure and safe to use. Below, we clearly and transparently break down how the MCP works with your data and the customer data you control, so you can decide what’s right for your business with confidence. The short version: We’ve applied best practices to secure the Kajabi MCP, and ensure that you control what your AI tool can and can’t do with it.Documentation Index
Fetch the complete documentation index at: https://help.kajabi.com/llms.txt
Use this file to discover all available pages before exploring further.
What is an MCP?
MCP stands for Model Context Protocol. It’s a shared standard that lets an AI tool, like Claude, ChatGPT, or Cursor, connect directly to another piece of software and do real work inside it. We built the Kajabi MCP so you can point an AI tool at your Kajabi account and have it make actual changes for you: drafting and editing emails, building out pages, updating products and offers, organizing contacts, and working across the other features you already use. That’s genuinely powerful, and it’s also why we take the security side seriously. Giving an AI tool the ability to make changes in your account means giving it access and reach, so we’ve done a lot of work to keep you in control of exactly what it can and can’t do. Here’s how that works.Decide what the AI can access through the MCP
When you connect an AI tool (Claude, ChatGPT, Cursor, or another), you approve each permission individually. Your connection starts read-only, and you grant access from there. If you would like the AI to draft your emails but never touch your contacts you may grant the one, and skip the other. The riskiest permissions, like editing your theme code or adding account users, stay off until you deliberately turn them on. Anything you haven’t granted isn’t available to that connection at all, so there’s nothing for the AI to work around.Review what the AI does with the MCP
You stay in control of the MCP at two points, and both are yours to set. First, by default, your AI tool asks for your permission before it takes any action (making a change, pulling data, anything). Claude, ChatGPT, and Cursor all offer an “auto-approve” mode you can turn on to stop those prompts, but we’d recommend against it, including when you’re working with the Kajabi MCP. Leaving approvals on means you’re always asked before anything happens. Second, not every action works the same way in the MCP. Many changes are saved as drafts for you to review and publish when you’re ready. Others take effect right away. You have a moment to check which kind it is before you approve an action.How we protect your account
We built in a lot of safeguards around the MCP. A few key ones are:- The MCP can’t exceed your own access: The MCP is limited to what your user role allows, and it’s locked to your account. It can’t reach anyone else’s. You can always set up a specific user account for the MCP to have more granular controls over its access.
- Automated disconnection: Remove any tool in one click under Settings → Integrations & Webhooks. Connections you stop using expire on their own.
- Protection against prompt injection: Content your customers enter is treated as information, not as instructions the AI should follow.
- It was independently tested: In addition to our own rigorous testing, we had the MCP put through independent security testing before launch and addressed what we found.