FlowMCP Server

FlowMCP Server

Provides LocalServer and RemoteServer implementations for running MCP servers locally via stdio or remotely via HTTP/SSE, with simple and advanced deployment options.

Category
访问服务器

README

FlowMCP Server

This repository provides two server implementations compatible with the FlowMCP framework:

  • 🖥 LocalServer — for local, stdio-based execution
  • 🌐 RemoteServer — for network-based usage via HTTP and SSE

Table of Contents


Quickstart

Deploy with DigitalOcean

An autodeploy is only available for a stateless server (streamableHTTP) .

Deploy to DO

🖥 Local Server

The LocalServer is designed for local workflows, using standard input/output streams. It is ideal for CLI tools, testing, and development environments.

✅ Features

  • Lightweight and dependency-free I/O via stdin/stdout
  • Fully supports FlowMCP.activateServerTools(...)
  • Uses StdioServerTransport

🚀 Example Usage

import { LocalServer } from 'flowmcp-server'
import { FlowMCP } from 'flowmcp'
import { SchemaImporter } from 'schemaimporter'

const schemaList = await SchemaImporter.get( { withSchema: true } )
const arrayOfSchemas = schemaList.map(({ schema }) => schema)

const { activationPayloads } = FlowMCP.prepareActivations({ arrayOfSchemas })

const localServer = new LocalServer({ silent: true })
localServer.addActivationPayloads({ activationPayloads })

await localServer.start()

🔧 Configuration

localServer.setConfig({
  overwrite: {
    serverDescription: {
      name: 'My Local Server',
      description: 'CLI test server',
      version: '1.2.2'
    }
  }
})

🌐 Remote Server

The RemoteServer provides HTTP-based access to FlowMCP schemas using various protocols. It is ideal for frontend apps, remote agents, and networked integrations.

✅ Features

  • Supports 2 transport protocols:

    • streamable (HTTP with stateless communication)
    • sse (Server-Sent Events)
  • Multiple routes and schemas can be activated

  • Easily configurable

🚀 Example Usage

import { RemoteServer } from 'flowmcp-server'
import { FlowMCP } from 'flowmcp'

const remoteServer = new RemoteServer({ silent: true })

// Define routes with their configuration
const arrayOfRoutes = [
  {
    routePath: '/api',
    protocol: 'sse',
  }
]

// Pre-assign schemas to routes
const objectOfSchemaArrays = {
  '/api': [...] // Your schemas here
}

// Prepare route activation payloads
const { routesActivationPayloads } = RemoteServer.prepareRoutesActivationPayloads({
  arrayOfRoutes,
  objectOfSchemaArrays,
  envObject: process.env
})

remoteServer.start({ routesActivationPayloads })

🔧 Configuration

remoteServer.setConfig({
  overwrite: {
    port: 8081,
    rootUrl: 'http://mydomain.com'
  }
})

📡 Supported Transport Protocols

Protocol Description
sse Server-Sent Events, persistent connection
streamable Stateless POST-based HTTP communication

🚀 Simple Deployment

The Deploy class provides a quick way to set up servers with command-line parameter support.

📝 Example Usage

import { Deploy } from 'flowmcp-server'

// Initialize with command-line arguments and schemas
const { serverType, app, mcps, events, argvs } = Deploy.init({
  argv: process.argv,
  processEnv: process.env,
  arrayOfSchemas: [...] // Your schemas
})

// Access parsed command-line arguments
console.log('Server Type:', serverType)     // 'local' or 'remote'
console.log('Parsed Args:', argvs)          // All CLI parameters
console.log('Express App:', app)            // Express.js app (remote) or McpServer (local)
console.log('MCPs:', mcps)                  // null for local, sessions object for remote
console.log('Events:', events)              // null for local, event emitter for remote

// Start the configured server
await Deploy.start()

🚀 Advanced Multi-Route Deployment

The DeployAdvanced class enables deployment of multiple routes with different schemas and protocols. Perfect for complex API setups.

🌟 Key Features

  • Multiple routes with independent schema sets
  • Pre-filtered schema assignment per route
  • Mixed transport protocols (SSE + HTTP)
  • Individual authentication per route

📝 Example Usage

import { DeployAdvanced } from 'flowmcp-server'

// Initialize the advanced deployment
const { serverType, app, mcps, events, server } = DeployAdvanced.init({ silent: true })

// Define routes with their configuration
const arrayOfRoutes = [
  {
    routePath: '/crypto',
    protocol: 'sse', 
  },
  {
    routePath: '/admin',
    protocol: 'streamable',
  }
]

// Pre-assign schemas to routes (user controls filtering)
const objectOfSchemaArrays = {
  '/crypto': [
    // Crypto-related schemas only
    coinGeckoSchema,
    deFiLlamaSchema
  ],
  '/admin': [
    // Admin-only schemas
    userManagementSchema,
    systemStatsSchema
  ]
}

// Start with pre-configured routes and schemas
DeployAdvanced.start({
  arrayOfRoutes,
  objectOfSchemaArrays, 
  envObject: process.env,
  rootUrl: 'https://api.example.com',
  port: 8080
})

// Optional: Access server internals for advanced customization
console.log('Server Type:', serverType)  // 'multipleRoutes'

// Express.js app - add custom middleware, routes, etc.
app.use('/health', (req, res) => res.json({ status: 'healthy' }))

// Monitor MCP sessions - track active connections per route
Object.entries(mcps).forEach(([route, { sessionIds }]) => {
  console.log(`Route ${route}: ${Object.keys(sessionIds).length} active sessions`)
})

// Event monitoring - listen to server events
events.on('sessionCreated', ({ protocol, routePath, sessionId }) => {
  console.log(`New ${protocol} session: ${sessionId} on ${routePath}`)
})

// Direct server access - modify configuration, add routes, etc.
server.setConfig({ overwrite: { port: 9000 } })

🔄 Migration from v1.3.x

OLD API (v1.3.x):

const routes = [{
  includeNamespaces: ['coingecko'],
  excludeNamespaces: ['debug'],
  activateTags: ['production'],
  routePath: '/crypto',
  protocol: 'sse',
}]

DeployAdvanced.start({
  routes,                    // ❌ Old parameter
  arrayOfSchemas: [...],     // ❌ Global array
  envObject: process.env
})

NEW API (v1.4.x):

const arrayOfRoutes = [{
  routePath: '/crypto',      // ✅ Simplified route
  protocol: 'sse',
}]

const objectOfSchemaArrays = {
  '/crypto': [...]           // ✅ Pre-filtered per route
}

DeployAdvanced.start({
  arrayOfRoutes,             // ✅ New parameter
  objectOfSchemaArrays,      // ✅ Route-specific schemas  
  envObject: process.env
})

🔧 Advanced Server Access

Both Deploy.init() and DeployAdvanced.init() return important objects that allow deep customization:

Object Deploy (Simple) DeployAdvanced Description
serverType 'local' or 'remote' 'multipleRoutes' Server configuration type
app Express app or McpServer Express app Server application instance
mcps null (local) or sessions object Sessions object Active MCP connections per route
events null (local) or EventEmitter EventEmitter Event system for monitoring
argvs Parsed CLI arguments null Command-line parameters (Deploy only)
server Not available RemoteServer instance Direct server access (DeployAdvanced only)

💡 Use Cases

  • Custom Middleware: Add authentication, logging, rate limiting via app
  • Connection Monitoring: Track active sessions via mcps and events
  • Health Checks: Add custom endpoints for monitoring
  • Configuration: Modify server settings via server (DeployAdvanced)
  • CLI Integration: Access parsed arguments via argvs (Deploy)

📌 Compatibility

  • FlowMCP Server version: 1.5.0
  • FlowMCP Schema spec version: 1.2.2

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选