Resources is an MCP protocol primitive that allows servers to expose content to AI agents in a structured, discoverable way — content that can be read or, in some cases, subscribed to for live updates. In the N3uron MCP Server, custom resources are files uploaded and stored directly inside the module configuration, making them available to any MCP-compatible client with resource access. The AI-Ready PV Demo Project includes one custom resource: the Xantrex GT500 Datasheet — the planning, installation, characteristics, and troubleshooting manual for the 500 kW grid-tied photovoltaic inverters modelled in the simulation.
Note:
Not all MCP clients fully support the Resources primitive. Some clients may not implement resource discovery or retrieval at all. Verify your client's capabilities before relying on resources in a workflow.
Embedded file resource: Xantrex GT500 Datasheet
The backup includes the Xantrex GT500 Datasheet embedded as a file resource. This is the inverter model that the demo simulation is built around.
Property | Value |
|---|---|
Name | Xantrex_GT500_Datasheet |
Title | Xantrex GT500 Datasheet |
Description | Xantrex GT500 Datasheet |
File | Embedded file — 184.45 KB (Xantrex GT500 MVX Planning and Installation Manual) |
.png)
How this resource enhances AI diagnostics
The Xantrex GT500 manual gives AI agents direct access to the inverter's technical specifications, electrical parameters, MPPT configuration, fault codes, and maintenance procedures. When an agent is investigating an underperformance finding or a recurring alarm, it can read the manual as context and cross-reference the operational data against the manufacturer's specifications — producing more precise and actionable recommendations grounded in the actual equipment rather than generic inverter knowledge.
The make_maintenance_plan prompt is specifically designed to use this: it instructs the AI model to cross-reference tool findings against any available documents before making recommendations, and to cite those references in the output. With the datasheet available, the model can reference specific sections — MPPT voltage thresholds, fault recovery procedures — rather than producing generic maintenance advice.
Tip:
The same pattern applies to any real deployment. Embedding datasheets, communication protocol specifications, wiring diagrams, or O&M procedures as resources gives AI agents the technical grounding needed to produce recommendations that are specific to the actual hardware — not generic. The more relevant context available to the agent, the more precise and trustworthy its output.
While most AI agent clients support direct file uploads, storing technical documents as MCP Server resources offers a meaningful advantage: the documents live alongside the data, inside the same node configuration, and are automatically available to every client that connects — without requiring each user or application to upload them separately. This makes the knowledge centralized, version-controlled with the node backup, and consistently accessible across all workflows and clients that use that MCP Server instance.
Adding or replacing embedded file resources
To add a new embedded file resource or replace the existing datasheet, navigate to MCP Server → Resources in the WebUI.
Navigate to MCP Server → Resources.
Select New Resource from the model menu to create a new entry, or select an existing resource to edit it.
Fill in Name, Title, and Description. Use a clear, descriptive Name — this is the identifier the AI agent uses to discover and request the resource.
Click the File field action menu (···) and select Select file to upload the document from your local machine.
Save the configuration. The resource is now embedded in the MCP Server and immediately available to any client token that has resource access enabled in the Access configuration.
Note:
Embedded file resources can be any file type, including binary formats such as PDF or images. However, when a binary file is delivered to an MCP client, the client is responsible for decoding it — and not all clients handle binary content.
To ensure the broadest compatibility and that the AI agent can read the content directly, it is recommended to convert documents to plain text before embedding them. In this demo, the Xantrex GT500 inverter manual was originally a PDF and was converted to a .txt file before being uploaded as a resource — making its content immediately readable by any MCP-compatible client without any additional decoding step.
Further details can be found in the MCP Server Resources.