MCP Server Boilerplate
shellsage-ai/mcp-server-boilerplate
Boilerplate production-ready para criar servidores MCP em TypeScript/Python
Ver no GitHubTemplates prontos para produção em TypeScript e Python para construir servidores Model Context Protocol (MCP). Fornece exemplos funcionais de tools, resources e prompts que seus agentes AI podem utilizar. Estrutura completa com configuração Claude Desktop incluída.
Instalação
npx convez-ai shellsage-ai/mcp-server-boilerplate1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Construir servidores MCP personalizados para agentes Claude
- Integrar ferramentas customizadas com Claude Desktop
- Desenvolver APIs que AI agents possam consumir nativamente
Veredito

Starter template sólido que acelera desenvolvimento de servidores MCP com exemplos de ferramentas, recursos e configuração. Ideal para quem quer conectar Claude a ferramentas customizadas sem começar do zero.
README
A tradução para o português deste README ainda está na fila.
MCP Server Boilerplate
Production-ready starter templates for building Model Context Protocol servers in TypeScript and Python.
Skip the boilerplate. Start building tools your AI agents can actually use.
Server Capabilities
This boilerplate ships with working examples of all three MCP primitives:
Tools
| Tool | Description | Parameters |
|---|---|---|
echo | Echo a message back to the caller | message (string, required) |
timestamp | Get the current UTC timestamp | (none) |
Resources
| URI | Description | MIME Type |
|---|---|---|
server://info | Server metadata (name, version, available tools) | application/json |
Prompts
No prompt templates are registered in the starter. The full kit includes prompt template patterns.
What's Included (Free)
TypeScript Quickstart
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
const server = new McpServer({
name: "my-mcp-server",
version: "1.0.0",
});
// Register a tool
server.tool("hello", { name: { type: "string" } }, async ({ name }) => ({
content: [{ type: "text", text: `Hello, ${name}!` }],
}));
// Connect via stdio
const transport = new StdioServerTransport();
await server.connect(transport);
Python Quickstart
from mcp.server import Server
from mcp.server.stdio import stdio_server
app = Server("my-mcp-server")
@app.tool()
async def hello(name: str) -> str:
"""Say hello to someone."""
return f"Hello, {name}!"
async def main():
async with stdio_server() as (read, write):
await app.run(read, write)
Claude Desktop Configuration
{
"mcpServers": {
"my-server": {
"command": "npx",
"args": ["tsx", "src/index.ts"]
}
}
}
Project Structure
my-mcp-server/
├── src/
│ └── index.ts # Server entry point
├── tools/
│ └── example.ts # Tool definitions
├── resources/
│ └── example.ts # Resource providers
├── package.json
├── tsconfig.json
└── claude_desktop_config.json
Getting Started
- Clone this repo
npm installnpm run build- Add to your Claude Desktop config
- Start building tools
Going Further
This free boilerplate gets you started. The MCP Server Boilerplate Kit ($49) includes:
- ✅ Full TypeScript + Python dual-language templates
- ✅ Docker containerization with multi-stage builds
- ✅ CI/CD pipeline (GitHub Actions) for automated testing & deployment
- ✅ SSE (Server-Sent Events) transport for web deployments
- ✅ 15+ pre-built tool examples (file ops, API calls, database queries)
- ✅ Resource and prompt template patterns
- ✅ Error handling, logging, and retry patterns
- ✅ Testing framework with mock MCP client
- ✅ Production deployment guide (Docker, systemd, cloud)
- ✅ Claude Desktop + Cursor + Windsurf integration configs
Resources
License
MIT — use this however you want.