
example-mcp-server-stdio
example-mcp-server-stdio
README
Calculator Learning Demo - STDIO Transport
<div align="center">
</div>
🎯 Overview
This repository demonstrates a learning-edition MCP calculator server using STDIO transport. It showcases the Model Context Protocol (MCP) SDK implementation with standard tools/list
and tools/call
methods, communicating via stdin
and stdout
using JSON-RPC messages.
This transport is the most performant and secure option for local inter-process communication (IPC).
Key Characteristics
- Transport Layer: Direct
stdin
/stdout
pipes between parent (client) and child (server) processes. - State Model: Ephemeral, in-process memory. All state (e.g., calculation history) is lost when the process exits.
- Latency: The lowest possible latency (microseconds), as it avoids all network stack overhead.
- Security: Extremely high due to OS-level process isolation. There is no network attack surface.
- Use Case: Ideal for high-performance, secure local tooling, such as command-line interfaces (CLIs), IDE plugins, and build-system integrations.
📊 Transport Comparison
This table compares the four primary MCP transport mechanisms demonstrated in the learning series. The implementation in this repository is highlighted.
Dimension | STDIO | SSE (Legacy) | Streamable HTTP (Stateful) | Streamable HTTP (Stateless) |
---|---|---|---|---|
Transport Layer | Local Pipes (stdin /stdout ) |
2 × HTTP endpoints (GET +POST ) |
Single HTTP endpoint /mcp |
Single HTTP endpoint /mcp |
Bidirectional Stream | ✅ Yes (full duplex) | ⚠️ Server→Client only | ✅ Yes (server push + client stream) | ✅ Yes (within each request) |
State Management | Ephemeral (Process Memory) | Ephemeral (Session Memory) | Persistent (Session State) | ❌ None (Stateless) |
Resumability | ❌ None | ❌ None | ✅ Yes (Last-Event-Id ) |
❌ None (by design) |
Scalability | ⚠️ Single Process | ✅ Multi-Client | ✅ Horizontal (Sticky Sessions) | ♾️ Infinite (Serverless) |
Security | 🔒 Process Isolation | 🌐 Network Exposed | 🌐 Network Exposed | 🌐 Network Exposed |
Ideal Use Case | ✅ CLI Tools, IDE Plugins | Legacy Web Apps | Enterprise APIs, Workflows | Serverless, Edge Functions |
📐 Architecture and Flow
The STDIO transport architecture is based on a parent-child process model. A client application spawns the MCP server as a child process and communicates with it by writing to the child's stdin
and reading from its stdout
.
sequenceDiagram
participant Client (Parent Process)
participant Server (Child Process)
Client->>Server: Writes JSON-RPC request to stdin stream
Server-->>Client: Writes JSON-RPC response to stdout stream
Note over Client,Server: Communication is full-duplex, concurrent, and newline-delimited.
MCP SDK Implementation
This repository contains a single, production-ready MCP server implementation:
dist/server.js
- Built with the official@modelcontextprotocol/sdk
- Uses standard MCP methods:
tools/list
,tools/call
,resources/list
,prompts/list
- High-level abstractions with
server.registerTool()
andserver.registerResource()
- Compatible with all MCP clients and registries (including Smithery)
✨ Feature Compliance
This server implements the complete MCP Latest Standard feature set for the learning edition.
Name | Status | Implementation |
---|---|---|
calculate |
Core ✅ | Basic arithmetic with optional streaming progress. |
batch_calculate |
Extended ✅ | Processes multiple calculations in a single request. |
advanced_calculate |
Extended ✅ | Factorial, logarithm, and combinatorics operations. |
demo_progress |
Extended ✅ | Demonstrates progress notifications over the stdout stream. |
explain-calculation |
Core ✅ | Returns a Markdown explanation of a calculation. |
generate-problems |
Core ✅ | Returns Markdown-formatted practice problems. |
calculator-tutor |
Core ✅ | Returns Markdown-formatted tutoring content. |
solve_math_problem |
Extended ✅ | Solves a math problem, may elicit input. |
explain_formula |
Extended ✅ | Provides an interactive formula explanation. |
calculator_assistant |
Extended ✅ | Offers interactive calculator assistance. |
calculator://constants |
Core ✅ | Resource for mathematical constants (π, e, φ, etc.). |
calculator://history/{id} |
Extended ✅ | Resource for the last 50 calculation results stored in memory. |
calculator://stats |
Extended ✅ | Resource for server uptime and request statistics. |
formulas://library |
Extended ✅ | Resource for a collection of mathematical formulas. |
🚀 Getting Started
Prerequisites
- Node.js (v18.x or higher)
- npm or yarn
Installation
# Clone the repository
git clone https://github.com/modelcontextprotocol/mcp-server-examples.git
cd mcp-server-examples/stdio
# Install dependencies
npm install
# Build the project
npm run build
Running the Server
The server is designed to be spawned by a client. You can run it directly to send it commands interactively.
# Run the MCP server
npm start
# Run directly with Node.js
node dist/server.js --stdio
# Run in development mode
npm run dev
Testing with MCP Inspector
Interact with the SDK-based server using the official MCP Inspector CLI. This command spawns the server and pipes its I/O to the inspector.
npx @modelcontextprotocol/inspector --cli "node dist/server.js --stdio"
📋 API Usage Examples
All requests use standard MCP protocol with JSON-RPC messages.
Standard MCP Protocol
The server implements the standard MCP SDK protocol:
# List available tools
→ {"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}
# Call a tool
→ {"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"calculate","arguments":{"a":7,"b":6,"op":"multiply"}}}
# Response
{"jsonrpc":"2.0","id":2,"result":{"content":[{"type":"text","text":"7 × 6 = 42"}]}}
Progress Demonstration
Progress notifications are sent as standard JSON-RPC notifications (no id
field) over stdout
.
# Request
→ {"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"demo_progress","arguments":{}}}
# Response Stream
{"jsonrpc":"2.0","method":"progress","params":{"progressToken":"progress_3","progress":20,"total":100}}
{"jsonrpc":"2.0","method":"progress","params":{"progressToken":"progress_3","progress":40,"total":100}}
{"jsonrpc":"2.0","method":"progress","params":{"progressToken":"progress_3","progress":60,"total":100}}
{"jsonrpc":"2.0","method":"progress","params":{"progressToken":"progress_3","progress":80,"total":100}}
{"jsonrpc":"2.0","method":"progress","params":{"progressToken":"progress_3","progress":100,"total":100}}
{"jsonrpc":"2.0","id":3,"result":{"content":[{"type":"text","text":"Progress demonstration completed"}]}}
🧠 State Management Model
Principle: State is ephemeral and strictly scoped to the lifetime of the server process.
- Mechanism: All state is held in standard JavaScript variables and
Map
objects within the Node.js process. - Stored Data:
- Calculation History: A
Map
stores the last 50 calculation results as a ring buffer. - Server Statistics: Variables track the process start time and total request count.
- In-flight Requests: The MCP SDK maintains a
Map
to track concurrent requests and route responses correctly.
- Calculation History: A
- Lifecycle: When the process exits for any reason, all in-memory state is irrevocably lost. Each new process starts with a clean slate. This is a fundamental and intentional design choice for this transport.
🛡️ Security Model
The STDIO transport provides the most secure environment of all MCP transports by leveraging operating system primitives.
- Process Isolation: The server runs in a separate memory space from the client, preventing any direct memory access or interference. The OS enforces this boundary.
- No Network Exposure: Communication is entirely via local IPC pipes. There are no open ports, making network-based attacks (e.g., CSRF, MitM, remote exploits) impossible.
- Input Validation: All incoming request parameters are rigorously validated by Zod schemas (defined in
src/types.ts
) to ensure type safety and prevent injection-style attacks. - Resource Limiting: The server enforces hard limits on batch sizes (
maxBatchSize: 100
) and history storage (maxHistorySize: 50
) to prevent resource exhaustion attacks. - Exit Code Signaling: The server uses standard Unix exit codes to signal its termination status to the parent process (e.g.,
0
for success,65
for data errors,70
for software errors), allowing the client to react appropriately.
🧪 Testing
This project includes a test suite that spawns the server as a child process to validate its I/O behavior.
# Run all tests defined in jest.config.js
npm test
# Run tests and generate a code coverage report
npm run test:coverage
# Run tests in watch mode for interactive development
npm run test:watch
📚 Official Resources
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.