resource

Resources are an, often overlooked, primitive of the MCP spec. You can think of them more or less as files that your MCP server can expose (in fact most MCP clients allow you to add them to the context with the same user interaction).

Differently from tools the LLM is only involved as a recipient of resources and doesn't autonomously pick them. Instead the user has to select the resources that it wish to include in the message which will then be sent as additional context for the LLM.

Basic API

You can register a resource by invoking the resource method on the server instance. The first argument is a configuration object and the second a handler that will be invoked whenever that resource is read by the MCP client.

		server.resource(
	{
		uri: 'mymcp://name-of-the-resource.json',
		name: 'your-resource',
		description: 'A description for the LLM',
		title: 'Your Resource',
	},
	(uri) => {
		return {
			contents: [
				{
					uri,
					text: 'Content of the resource',
					mimeType: 'text/plain', // this is optional
				},
			],
		};
	},
);
	

uri, name and description are the only required properties (you can also specify a title for a human readable title but that's optional) of the configuration object. The return value of the handler must be an object with a contents property which is an array of one or more contents (you can refer to the MCP spec to know the possible return values).

The uri must conform to the RFC3986 standard and it can be wether a common uri (like https or git) or it could be a custom one.

enabled function

One pattern that is quite common in every software is having a different feature-set based on some flag or the status of some user. You could technically create a new instance of the McpServer for each request and conditionally add a resource but to facilitate the process tmcp exposes an enabled property on the configuration object. The property is a function that returns a boolean and, as you might have guessed, allows you to include a specific resource in the list of resources conditionally. Within the function you have access to the context so you can make decisions based on the client capabilities, the client info or even just reading a feature flag in the db to do A/B testing or to allow your admin to turn on or off a resource without a re-deploy.

		server.resource(
	{
		uri: 'mymcp://name-of-the-resource.json',
		name: 'your-resource',
		description: 'A description for the LLM',
		title: 'Your Resource',
		enabled() {
			return server.ctx.sessionInfo?.clientInfo?.name !== 'codex';
		},
	},
	(uri) => {
		return {
			contents: [
				{
					uri,
					text: 'Content of the resource',
					mimeType: 'text/plain', // this is optional
				},
			],
		};
	},
);
	

Icons

To allow the users to to understand what an MCP server is about at a glance the MCP spec allows you to include a set of icons for each resource. Obviously tmcp allows you to specify those too using the icons property of the configuration object.

		server.resource(
	{
		uri: 'mymcp://name-of-the-resource.json',
		name: 'your-resource',
		description: 'A description for the LLM',
		title: 'Your Resource',
		icons: [
			{
				src: 'https://dantemcp.com/date.png',
			},
			{
				src: 'data:image/png;base64,...',
			},
		],
	},
	(uri) => {
		return {
			contents: [
				{
					uri,
					text: 'Content of the resource',
					mimeType: 'text/plain', // this is optional
				},
			],
		};
	},
);
	

Dynamic Resources

The resource method is only used to add static resources to your MCP server, to learn how to add dynamic resources please check out the template documentation.