> For the complete documentation index, see [llms.txt](https://docs.struct.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.struct.com/tutorials/guides/how-to-use-mcp/how-to-create-mcp-configuration-and-connect-to-ai-assistant.md). # How to create MCP configuration and connect to AI assistant Set up a secure connection between your Struct PIM tenant and your AI assistant. ## Create your MCP configuration To create an MCP configuration, go into the "Settings"-tab from the navigation bar, then locate the "Integration"-section and click "MCP".

Here you can manage your existing MCP configurations, including viewing and deleting configurations. To create a new configuration, click "+ Create MCP configuration"-button in the top right corner.

This will open the MCP configuration page. ### Configure your MCP connection On the MCP configuration page, give your configuration a required name and set the permissions that should be available to the connected AI assistant.

The permissions determine which data and actions the assistant can access in your Struct PIM tenant. Only grant the permissions required for the intended use. Once you have configured the permissions, save the MCP configuration. Struct will generate an MCP API key for the configuration.

{% hint style="warning" %} Treat the MCP API key like a password. Do not share it publicly or include it in source control. You will need it when configuring your AI assistant. {% endhint %} ## Connect your AI assistant Once your MCP configuration has been created, you can connect it to your preferred AI assistant. Each assistant uses the same Struct MCP endpoint and the API key generated above. ### Claude Desktop Open Settings > Developer, and under the "Local MCP servers"-section, click the "Edit Config"-button.

You will be shown the config file in File Explorer (Finder on macOS), where you can edit it in Notepad or a similar text editor.

In the config file you need to add the `struct-platform` entry. It defines how Claude connects to the Struct PIM MCP endpoint. {% hint style="warning" %} Both configurations require Node.js to be installed, as it provides `npx`. {% endhint %} Paste the block right after the first `{` in the file, and remember to add a comma after the block’s closing `}`. If the file already contains `"mcpServers"`, don’t paste a new one. Instead, add only the `"struct-platform": { ... }` part inside the existing `"mcpServers"`. **MacOS** ```json "mcpServers": { "struct-platform": { "command": "npx", "args": [ "-y", "mcp-remote", "https://TENANT.mcp.struct.com/mcp", "--header", "Authorization: Bearer mcp_" ] } }, ``` **Windows** ```json "mcpServers": { "struct-platform": { "command": "cmd", "args": [ "/c", "npx", "-y", "mcp-remote", "https://TENANT.mcp.struct.com/mcp", "--header", "Authorization: Bearer mcp_" ] } }, ``` {% hint style="warning" %} Replace `mcp_` with the key from your MCP configuration and `TENANT` with your actual tenant. {% endhint %} After saving the configuration, restart Claude Desktop. The `struct-platform` server should appear under Settings > Developer with a running status. #### Confirm the connection Once your assistant has been configured, ask it something that requires live data from your Struct PIM tenant. For example:

Claude Desktop - Example of MCP chat response

If the assistant returns live data from Struct PIM, your MCP connection is working correctly. If you receive an authentication error, check that the Bearer token contains the correct MCP API key and that the permissions assigned to the configuration allow access to the data you requested. ### Claude Code Add the server from the terminal Run the following command in PowerShell or your terminal: ```bash claude mcp add --transport http --scope user struct-platform \ https://TENANT.mcp.struct.com/mcp \ --header "Authorization: Bearer mcp_" ``` {% hint style="warning" %} Replace `mcp_` with the key from your MCP configuration and `TENANT` with your actual tenant. The `--scope user` flag makes the server available across all your projects. {% endhint %} ### ChatGPT Connect the ChatGPT desktop app using the app's connection form. This guide uses the **Headers** fields to send your Struct access token. Open Settings from the profile menu at the bottom left of the app. In Settings, go to Integrations -> Plugins. The page is headed Plugins and shows the tabs Plugins, Apps, and MCPs. *The desktop app manages MCP connections under Plugins. There is no separate MCP servers item in this version's Settings sidebar.*

At the top right of the Plugins settings page, select Add -> Add MCP server.

Complete the form as follows: | Field | What to enter | | -------- | ---------------------------------------- | | **Name** | `Struct MCP` | | **Type** | Streamable HTTP | | **URL** | | Use the exact URL from your Struct administrator.

Leave Bearer token env var empty. Under Headers, complete the key/value row with: | Header field | Value | | ------------ | --------------------------------- | | **Key** | `Authorization` | | **Value** | `Bearer YOUR_STRUCT_ACCESS_TOKEN` | Replace `YOUR_STRUCT_ACCESS_TOKEN` with your actual token. Keep the word `Bearer` followed by one space before the token. Do not add quotation marks in the UI fields. Scroll to the bottom of the form and select Save. Check the MCPs tab in Plugins settings and confirm that your connection is enabled. If its tools do not appear in a chat, fully quit and reopen the desktop app, then return to the chat. {% hint style="warning" %} Custom MCP servers are enabled on Business/Enterprise workspaces with developer mode. {% endhint %} ### Copilot studio Copilot Studio connects to the Struct MCP server as an MCP tool on your agent. Open your agent, and in the right-hand panel, click the “+”-button next to Tools.

In the “Add a tool” dialog, click the “Add”-button and select MCP (Model Context Protocol).

Fill in the server details in the wizard: * **Server name:** Struct PIM * **Server description:** a short description of what the server does, e.g. “Access and manage product data in Struct PIM” * **Server URL:** `https://TENANT.mcp.struct.com/mcp` * **Authentication:** API key * **Parameter type:** Header * **Header name:** `Authorization` * **Key value:** `Bearer mcp_` {% hint style="warning" %} Remember to include `Bearer` in front of the key. Replace `mcp_` with the key from your MCP configuration and `TENANT` with your actual tenant. {% endhint %} Click the “Add”-button. Under "Connection", open the dropdown and click “Create new connection”.

In the Authorization field, enter your MCP API key in the format `Bearer mcp_`, then click the “Create”-button. Click the “Add to agent”-button to add the tool to your agent.

The Struct PIM tool should now appear under Tools. Click it to see the available tools.

By default, each user has to sign in with their own connection. To let all users share the same API key, set Authentication to Shared account in the tool settings.

## Struct MCP tools overview More information about Struct MCP and and its tools, can be found here: {% file src="/files/b6HlNoYbbKYv9rSuXvDK" %} --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.struct.com/tutorials/guides/how-to-use-mcp/how-to-create-mcp-configuration-and-connect-to-ai-assistant.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.