About MultiServerMCP
MultiServerMCP is an SSE-based framework for serving multiple client connections from one MCP server. It includes URL-based tool and permission grouping, heartbeat handling, and session context access.
Key features
- 01Multiple SSE client connections
- 02URL-based tool and permission groups
- 03Connection heartbeat mechanism
- 04Session and client context management
- 05Simplified tool registration
Use cases
- 01Serving multiple clients from one server
- 02Grouping tools by URL path
- 03Applying permissions to tool groups
- 04Reading request data by session ID
MCP support
Implements an MCP server over SSE and supports registering tools, resources, and prompts for connected MCP clients.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEMultiServerMCP
A multi-connection MCP server framework based on SSE for long-connection communication. Provides context and functional extension capabilities for AI assistants. Currently only supports SSE mode.
🚀 MultiServerMCP — Enable AI assistants with extended capabilities through multi-connection server framework!
MultiServerMCP is designed to provide a robust server framework that supports multiple client connections to a single server, allowing for efficient tool and permission management for AI assistants.
🌟 Main Features
- Support for multiple client connections to a single server in SSE mode
- Tool and permission management by URL grouping
- Built-in heartbeat mechanism to ensure connection stability
- Simplified tool registration process
- Full compatibility with MCP protocol
- Global session management with
SessionManagerto access client context anywhere - Convenient global functions to retrieve request parameters and client context by sessionId
📦 Installation
npm install multi-server-mcp
# or
yarn add multi-server-mcp
# or
pnpm add multi-server-mcp
🚀 Quick Start
Complete Server Example
import { z } from "zod";
import { MultiServerMCP } from "multi-server-mcp";
import { ResourceTemplate } from "@modelcontextprotocol/sdk/server/mcp.js";
import { getSseReqQuery } from "multi-server-mcp";
// Create MCP server with URL grouping enabled
const server = new MultiServerMCP(
{
name: "multi server",
version: "1.0.0",
},
{
enableUrlGroups: true,
}
);
// Register a calculator tool that uses request parameters
server.tool("calc/add", { a: z.number(), b: z.number() }, async ({ a, b }, extra) => {
console.log(extra);
// Get query parameters from the SSE connection
const reqQuery = getSseReqQuery(extra.sessionId as string);
console.log(reqQuery);
return {
content: [{ type: "text", text: String(a + b) }],
};
});
// Register a resource template
server.resource(
"greeting",
new ResourceTemplate("greeting://{name}", { list: undefined }),
async (uri, { name }) => ({
contents: [
{
uri: uri.href,
text: `Hello, ${name}!`,
},
],
})
);
// Start the server
server
.start({
transportType: "sse",
})
.then(() => {
console.log(`Server started`);
})
.catch((error) => {
console.error(`Server start failed: ${error}`);
process.exit(1);
});
Basic Usage
import { MultiServerMCP } from 'multi-server-mcp';
// Create MCP server instance
const server = new MultiServerMCP({
name: 'my-mcp-server',
version: '1.0.0'
});
// Register a tool
server.tool('my-tool', async () => {
return {
content: [{ type: 'text', text: 'Tool executed successfully!' }]
};
});
// Start the server
server.start({
transportType: 'sse',
sse: {
port: 3000,
endpoint: '/mcp',
messagesEndpoint: '/mcp-messages'
}
});
Tool Grouping and Permission Management
// Register a tool with grouping
server.tool('group1/group2/my-tool', async () => {
return {
content: [{ type: 'text', text: 'Tool executed successfully!' }]
};
});
// Start server with permission management
server.start({
transportType: 'sse',
enableUrlGroups: true
});
Accessing Client Context Globally
import { getSseReqQuery, getClientContextBySessionId } from 'multi-server-mcp';
import { z } from "zod";
// Register a tool that uses request parameters
server.tool("calc/add", { a: z.number(), b: z.number() }, async ({ a, b }, extra) => {
// Get request parameters from the SSE connection
const reqQuery = getSseReqQuery(extra.sessionId as string);
console.log('Query parameters:', reqQuery);
return {
content: [{ type: "text", text: String(a + b) }],
};
});
// In any part of your code, you can also access client's request parameters
function myFunction(sessionId: string) {
// Get query parameters from the SSE connection request
const reqQuery = getSseReqQuery(sessionId);
console.log('Client query parameters:', reqQuery);
// Get the full client context if needed
const clientContext = getClientContextBySessionId(sessionId);
if (clientContext) {
console.log('Client URL groups:', clientContext.urlGroups);
// Access any other client context properties
}
}
🔧 Development Guide
Currently only SSE mode is supported.
Supported Tool Types
- Regular tools:
server.tool() - Resource tools:
server.resource() - Prompt tools:
server.prompt()
Session Management
The SessionManager provides a singleton instance to manage client sessions:
SessionManager.getInstance(): Get the singleton instanceSessionManager.getInstance().getClientContext(sessionId): Get client context by sessionIdSessionManager.getInstance().getReqQuery(sessionId): Get request query parameters by sessionIdSessionManager.getInstance().getSessionCount(): Get the count of active sessions
For convenience, the following global functions are provided:
getSseReqQuery(sessionId): Get request query parameters by sessionIdgetClientContextBySessionId(sessionId): Get client context by sessionId
🤝 Contributing
Contributions, issues, and feature requests are welcome!
📄 License
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.