FlowMCP transforms your schemas into production-ready MCP servers that AI applications can use immediately.
Basic MCP Server Setup
Create an MCP server that transforms your schemas into AI-accessible tools:
import { FlowMCP } from './src/index.mjs'
import { Server } from '@modelcontextprotocol/sdk/server/index.js'
// Your schemas become AI tools
const schemas = [
githubSchema // User profiles, repositories
]
// Create MCP server instance
const server = new Server({
name: 'my-flowmcp-server',
version: '1.0.0'
}, {
capabilities: { tools: {} }
})
// Server parameters for API authentication
const serverParams = {
GITHUB_TOKEN: process.env.GITHUB_TOKEN
}
// Transform schemas into MCP tools
const { mcpTools } = FlowMCP.activateServerTools({
server,
schema: githubSchema,
serverParams,
validate: true,
silent: false
})
console.log('Activated tools:', Object.keys(mcpTools))
// Output: ['github__getUser']
Multiple Schema Activation
// Activate all schemas at once
schemas.forEach(schema => {
FlowMCP.activateServerTools({
server,
schema,
serverParams,
validate: true
})
})
// Result: AI tools from your schemas
Claude Desktop Integration
Step 1: Create Your Server Script (flowmcp-server.js)
import { FlowMCP } from './src/index.mjs'
import { Server } from '@modelcontextprotocol/sdk/server/index.js'
const server = new Server({
name: 'claude-flowmcp-server',
version: '1.0.0'
}, {
capabilities: { tools: {} }
})
const serverParams = {
GITHUB_TOKEN: process.env.GITHUB_TOKEN
}
// Activate your schemas
const { mcpTools } = FlowMCP.activateServerTools({
server,
schema: githubSchema,
serverParams
})
console.log('✅ FlowMCP server ready for Claude Desktop')
console.log(`📡 Active tools: ${Object.keys(mcpTools).length}`)
claude_desktop_config.json:
{
"mcpServers": {
"flowmcp": {
"command": "node",
"args": ["/absolute/path/to/flowmcp-server.js"],
"env": {
"GITHUB_TOKEN": "your-github-token"
}
}
}
}
- Restart Claude Desktop
- Ask: “What tools do you have access to?”
- Test with: “Get the GitHub profile for octocat”
Production Configuration
// Production-ready server with security
const server = new Server({
name: 'production-flowmcp',
version: '1.0.0'
}, {
capabilities: { tools: {} }
})
// Enhanced server parameters
const serverParams = {
GITHUB_TOKEN: process.env.GITHUB_TOKEN
}
// Schema activation
FlowMCP.activateServerTools({
server,
schema: githubSchema,
serverParams,
validate: true
})
console.log('📊 Production server ready')
Advanced Configuration
Schema Filtering
Filter which schemas to activate based on environment or user needs:
// Filter schemas by namespace
const { filteredArrayOfSchemas } = FlowMCP.filterArrayOfSchemas({
arrayOfSchemas: allSchemas,
includeNamespaces: ['github'],
excludeNamespaces: [],
activateTags: ['production']
})
// Activate only filtered schemas
filteredArrayOfSchemas.forEach(schema => {
FlowMCP.activateServerTools({
server,
schema,
serverParams
})
})
Environment-Specific Deployment
const environment = process.env.NODE_ENV || 'development'
const config = {
development: {
validate: true,
silent: false,
includeNamespaces: ['github']
},
production: {
validate: true,
silent: true,
includeNamespaces: ['github']
}
}
const { includeNamespaces } = config[environment]
const { filteredArrayOfSchemas } = FlowMCP.filterArrayOfSchemas({
arrayOfSchemas: allSchemas,
includeNamespaces
})
console.log(`📦 ${environment} deployment with ${filteredArrayOfSchemas.length} schemas`)
Testing Your Server
Schema Validation Test
async function testServerSetup() {
console.log('🧪 Testing FlowMCP server setup...')
// Test schema validation
const validation = FlowMCP.validateSchema({ schema: githubSchema })
if (!validation.status) {
console.error('❌ Schema validation failed:', validation.messages)
return false
}
console.log('✅ Schema validation passed')
// Test API connectivity
const result = await FlowMCP.fetch({
schema: githubSchema,
userParams: { username: 'octocat' },
serverParams: { GITHUB_TOKEN: process.env.GITHUB_TOKEN },
routeName: 'getUser'
})
if (result.status) {
console.log('✅ API connectivity working')
console.log('🎉 Server setup complete!')
return true
} else {
console.error('❌ API test failed:', result.messages)
return false
}
}
testServerSetup()
async function testPerformance() {
console.log('⚡ Running performance tests...')
const startTime = Date.now()
const testRequests = 5
const promises = Array.from({ length: testRequests }, () =>
FlowMCP.fetch({
schema: githubSchema,
userParams: { username: 'octocat' },
serverParams: { GITHUB_TOKEN: process.env.GITHUB_TOKEN },
routeName: 'getUser'
})
)
await Promise.all(promises)
const averageTime = (Date.now() - startTime) / testRequests
console.log(`📊 Average response time: ${averageTime.toFixed(2)}ms`)
if (averageTime > 5000) {
console.warn('⚠️ Consider optimization for response times >5s')
}
}
Deployment Strategies
Local Development
# Simple development server
node server.mjs
# Development with environment variables
GITHUB_TOKEN=your-token NODE_ENV=development node server.mjs
Production Deployment
# Production deployment
node server.mjs
Docker Configuration
Troubleshooting
Common Issues
Server Won’t Start
- Validate schemas:
node -e "console.log(FlowMCP.validateSchema({schema}))"
- Verify dependencies:
npm list
Claude Desktop Integration
- Use absolute paths in configuration
- Restart Claude Desktop after config changes
- Check environment variables are set correctly
API Connectivity
- Test external APIs directly:
curl https://api.github.com/users/octocat
- Verify API keys are valid and not expired
- Check network connectivity and firewall settings
Health Monitoring
Next Steps
Expand Your Server: Add more schemas from the Schema Library to increase your server’s capabilities.
Production Optimization: Implement caching, rate limiting, and monitoring for production deployments.
Advanced Integration: Explore the Architecture Guide for sophisticated deployment patterns.
Success! You now have a working MCP server that transforms your schemas into AI-accessible tools. Your server can handle everything from simple API calls to complex multi-source data integration.
