What Inriver’s MCP means for the AI tools connected to your PIM

What does MCP do, and who does it apply to?

Getting AI tools to read accurately from your product data management setup has traditionally required a custom integration for each new connection, and for most teams, that overhead adds up quickly. 

MCP works as a wrapper around the Inriver REST API, exposing your PIM data through a protocol that AI tools like Claude, VSCode, and Cursor are already built to consume. Previously, each new AI tool your team wanted to connect to your PIM meant building and maintaining a separate integration. MCP replaces that with a consistent entry point, so your team spends less time on integration work and more time on what those tools are actually meant to do.

Where MCP fits into your setup depends largely on your product data readiness. If your AI tools are already running but pulling from scattered or incomplete sources, MCP gives them a more reliable, direct connection to your PIM, which directly affects the quality of outputs those tools return. The shift toward agentic B2B commerce is making that kind of structured data access more consequential, and the MCP vs UCP breakdown covers how MCP compares to other connectivity approaches if you want to weigh your options.

The two MCP endpoints and what each one does

If your team is connecting a PIM to AI tools, you typically run into two distinct needs: access to live product data for consumer-facing tools, and coding support for developers building or extending integrations. Inriver’s MCP addresses both through two separate endpoints, and knowing which one applies to your use case helps you set the right priorities for setup.

1. Query Manager

The Query Manager handles data access, enabling AI tools to query and fetch entity data from your PIM using structured criteria. If you want an LLM, an AI recommendation engine, or an analytics system pulling live product information from Inriver, this is the endpoint to configure first. 

Connecting those tools through the Query Manager gives them the context they need to return accurate, brand-aligned results, which matters most when those tools are customer-facing. 

Authentication requires a valid REST API key in the X-inRiver-APIKey header, and the user associated with that key needs the MCPUser permission assigned in Control Center. How well those tools perform will also depend on how thoroughly you’ve handled product data validation before they start pulling from your PIM. 

2. Code Writer

The Code Writer endpoint is aimed at your development team and lets them write and test REST API scripts and functions in supported environments, which is the more practical option if you’re actively building or extending Inriver integrations. 

Unlike the Query Manager, it requires no authentication, so your developers can get to work without waiting on API key setup or permission assignments, and your team gets a faster path to building new AI-driven capabilities on top of your PIM data.

Here’s a quick side-by-side of both endpoints to help you decide where to start:

What tools can you connect to your Inriver PIM using MCP? 

Inriver’s MCP currently works with three tools: Claude Desktop, Visual Studio Code (VSCode), and Cursor. If your team already operates in any of these environments, you can connect your PIM without rebuilding your existing workflow.

1. Claude Desktop

Claude Desktop supports MCP via a local JSON config file, with the location varying by operating system: on macOS, it’s in ~/Library/Application Support/Claude/, and on Windows, it’s in %APPDATA%\Claude\. MCP is available on Pro, Max, Team, and Enterprise plans, so confirm your plan before starting setup.

2. Visual Studio Code

VSCode offers two setup paths: a one-click installation link or manual configuration via an mcp.json file in your project folder. Both Query Manager and Code Writer are available via either method.

3. Cursor

Cursor supports one-click installation for both endpoints, though the Query Manager install uses a placeholder for the API key, so you’ll need to edit the config manually and insert your actual key before it works.

If your team is considering manual versus automated syndication methods, connecting these tools to your PIM through MCP increases automation in product content workflows without needing a complete rebuild of your existing system.

What are the MCP setup requirements?

Most of what the MCP setup requires is likely already in place if your team works with the Inriver REST API, since MCP uses the same authentication. Getting MCP configured requires three things before you start:

• A valid REST API key for your Inriver environment, passed in the X-inRiver-APIKey header
• The MCPUser permission assigned to the relevant user’s role in Control Center
• NodeJS and NPM installed on the machine running the connection

The MCPUser permission is the only requirement specific to MCP and must be assigned by your Inriver administrator directly in Control Center. NodeJS and NPM also need to be verified on the machine running the connection, since the setup process depends on both.

One thing worth noting for Claude Desktop specifically: the config file needs to be edited manually, and its location differs between macOS and Windows, so your team should confirm the correct path before making changes. 

If you’ve already worked through your content onboarding process, you’ll be better positioned to get useful outputs from MCP-connected tools right away, since the quality of data already in your PIM directly affects what those tools return.

What use cases does Inriver MCP support? 

The use cases MCP enables fall into two categories, depending on the endpoint your team is working with: consumer-facing AI applications that need live product data, and development tasks that involve building or extending API integrations.

Query Manager use cases

AI tools connected through the Query Manager can query your PIM directly, which makes it the right endpoint for applications where your product data needs to be accurate and current to function well:

• AI chatbots that need to pull accurate, up-to-date product information to support shoppers in real time without relying on static exports or manually maintained data sources
• Recommendation engines and analytics systems that need direct access to your PIM so the product data feeding those tools stays current without additional sync work
• Content generation and data validation workflows where AI tools draw from your actual PIM data rather than a separate copy of it
• Customer support tools that need reliable product information to respond accurately

Code Writer use cases

If your team is building toward more automated product content workflows will find that the endpoint reduces the development effort involved in connecting AI solutions to Inriver, giving your team and partners the tools to build new AI-driven capabilities faster:

• Writing and testing REST API interaction code within your working environment
• Building or extending Inriver integrations more efficiently without switching between tools
• Developing new AI-driven capabilities on top of your PIM data, particularly when paired with AI-powered data enrichment processes

Get your AI tools connected to your PIM system

MCP gives your AI tools a direct, structured path into your product data without custom integrations sitting in between, and the setup process is more straightforward than it might seem if your team already works with the Inriver REST API. The three prerequisites: your API key, the MCPUser permission, and NodeJS, are all you need before you start configuring your tool of choice.

If you want to see how MCP works in your specific setup or have questions about getting started, contact us for fast, tailored answers.