Connecting AI
Overview
iVendNext AI Gateway works by establishing a secure, authenticated connection between your iVendNext instance and an AI assistant — such as Claude Desktop or ChatGPT. Once this connection is in place, users can begin asking questions about their iVendNext data immediately, without any additional configuration.
This article explains how that connection works, what is required to set it up, and what users can expect once it is live.
What You Need Before You Start
Setting up iVendNext AI Gateway requires the following:
An active iVendNext instance with the AI Gateway module installed and enabled. Your system administrator will confirm whether the module is available on your site.
Administrator access for initial setup — the first step of the setup process must be completed by a user with the Administrator role in iVendNext. The Administrator enables the Gateway in iVendNext Settings → Assistant Core Settings and provides the MCP Endpoint URL to users who need to connect.
An MCP-compatible AI assistant — this is the tool your team will use to ask questions. iVendNext AI Gateway is compatible with Claude Desktop (Anthropic), ChatGPT (OpenAI), and any other AI assistant that supports the MCP connector standard.
Your own AI assistant subscription — iVendNext AI Gateway does not include a subscription to any AI assistant. You must have your own active subscription to a compatible AI assistant, obtained directly from the respective provider. The cost of that subscription is not covered by your iVendNext AI Gateway licence.
An iVendNext user account with the appropriate roles and permissions. The Gateway enforces iVendNext's existing permission model, so the user connecting the AI assistant must have a valid iVendNext login. What the AI can see and do is determined entirely by that user's roles — nothing more, nothing less.
The Connection Process
Connecting your AI assistant to iVendNext AI Gateway is a one-time setup. The steps below describe the process for Claude Desktop; the process is equivalent for other MCP-compatible assistants.
Step 1 — Obtain the MCP Endpoint URL
An Administrator must first enable the Gateway in iVendNext by navigating to Settings → Assistant Core Settings, checking the Enable Assistant Core checkbox, and saving. Once enabled, the MCP Endpoint URL is displayed on the General tab of that settings page. This URL is unique to your iVendNext instance and is what your AI assistant uses to communicate with the Gateway. The Administrator copies this URL and shares it with users who need to connect.
Step 2 — Add iVendNext as a Connector in Your AI Assistant
In Claude Desktop, navigate to Settings and then Connectors. Select Add Custom Connector, paste the MCP Endpoint URL provided by your administrator, and click Add. The AI assistant will register iVendNext AI Gateway as a new data connection.
For ChatGPT and other MCP-compatible tools, the equivalent option is typically found under Integrations or Connected Apps. The process follows the same pattern: paste the endpoint URL and confirm the connection.
Step 3 — Authenticate with Your iVendNext Account
Once the connector is added, you will be prompted to log in. Click Connect and you will be redirected to your iVendNext login screen. Enter your iVendNext credentials — the same username and password you use to access the application — and authorise the connection.
Step 4 — Confirm the Connection
Once authenticated, the AI assistant will confirm that iVendNext is connected. You are now ready to begin asking questions.
Understanding the Authentication Model
iVendNext AI Gateway uses OAuth 2.0 with PKCE — an industry-standard authentication protocol — to manage connections between AI assistants and iVendNext. In plain terms, this means:
Your password is never shared with the AI assistant. The AI tool never sees your iVendNext credentials. Authentication happens directly between your browser and iVendNext, and the AI assistant receives only a secure access token that it uses to make requests on your behalf.
The connection is scoped to your account. Every request the AI makes to iVendNext is made under your user identity. This is not a system-level connection with elevated privileges — it is your personal account, accessed by the AI assistant with your authorisation.
You can revoke access at any time. If you no longer want the AI assistant to have access to your iVendNext data, your administrator can revoke the connection from within iVendNext. The AI assistant will immediately lose access.
What Happens After Connection
Once connected, the AI assistant has access to iVendNext through the Gateway and can begin responding to questions about your business data. There is no additional training required and no prompts to configure.
It is worth noting that the connection persists between sessions. Once you have authenticated, you do not need to reconnect each time you open your AI assistant — unless your administrator resets or revokes the connection, or your session token expires.
Multiple Users, One Gateway
iVendNext AI Gateway supports multiple users connecting their own AI assistants to the same iVendNext instance. Each user authenticates independently with their own iVendNext account, and the Gateway enforces their individual permissions on every request.
This means a store manager and a managing director can both use the same Gateway installation to ask questions about the business — but the store manager will only see data within their permitted scope, while the MD sees data consistent with their broader access level. The Gateway does not grant any user access beyond what iVendNext already permits them.
Supported AI Assistants
iVendNext AI Gateway is compatible with any AI assistant that supports the MCP (Model Context Protocol) connector standard. At the time of writing, this includes:
Claude Desktop (Anthropic) — desktop application for Windows and Mac
ChatGPT (OpenAI) — web and desktop
Other MCP-compatible tools — any AI assistant that supports custom MCP connectors
As the MCP standard is adopted more widely across the AI industry, the number of compatible tools will grow. iVendNext AI Gateway is designed to work with any compliant client without requiring updates to the Gateway itself.
Related Articles
Automation and AI
Overview The iVendNext Node for n8n opens two distinct capabilities beyond basic data transfer: it enables complex multi-step automation patterns that combine real-time event reactions with scheduled batch jobs, and it makes iVendNext data available ...AI Gateway
What Is iVendNext AI Gateway? iVendNext AI Gateway is a module that connects your iVendNext instance directly to AI assistants — tools like Claude Desktop and ChatGPT that your team may already be using. Once connected, anyone in your organisation ...iVendNext AI Copilot - Appendix
Limits, defaults & glossary Default limits at a glance Setting Default Notes Requests per minute, per user 10 Configurable by administrator Requests per day, per user 200 Configurable by administrator Actions per question 15 Multi-step questions may ...Security & Audit
Overview When an AI assistant is given access to live business data, the natural question for any responsible organisation is: how is that access controlled, and what oversight exists? iVendNext AI Gateway is designed with two principles at its core ...iVendNext AI Copilot - Troubleshooting
If something does not look right Most issues resolve in a step or two. Work down this list before escalating. What you see What to do The Copilot button is missing Confirm Copilot is enabled for your environment and that you are signed in. Refresh ...