Connecting MCP servers to major AI platforms unlocks their full potential. This guide demonstrates integrating MCP servers with Azure AI Foundry (Microsoft Foundry) and Claude, the two leading platforms that provide first-class MCP support. These integrations enable your custom tools and data sources to work seamlessly with enterprise AI applications.
Microsoft Foundry represents Microsoft’s unified AI platform, combining model access, agent orchestration, and enterprise features. Claude Desktop and Claude Code from Anthropic pioneered MCP adoption with native protocol support. Both platforms enable sophisticated AI workflows powered by your custom MCP servers.
Integrating MCP Servers with Azure AI Foundry
Microsoft Foundry (formerly Azure AI Foundry) integrated MCP support in May 2025, bringing the protocol to Azure’s enterprise AI platform. Foundry Agent Service supports MCP tools natively, allowing agents to discover and invoke your custom MCP servers alongside built-in Azure services.
Register your MCP server in Foundry by providing the server URL, authentication configuration, and tool metadata. Foundry handles capability negotiation, tool discovery, and execution coordination. Your MCP server appears as a native tool source within Foundry Agent Service.
# Python: Configure MCP server in Azure AI Foundry
from azure.ai.projects import AIProjectClient
from azure.identity import DefaultAzureCredential
# Initialize Foundry client
project = AIProjectClient(
endpoint="https://your-foundry.ai.azure.com/api/projects/project-name",
credential=DefaultAzureCredential()
)
# Register MCP server as tool source
mcp_config = {
"type": "mcp",
"name": "custom-data-server",
"url": "https://your-mcp-server.azurewebsites.net/mcp",
"authentication": {
"type": "oauth2",
"token_endpoint": "https://auth.example.com/oauth/token",
"client_id": "your-client-id"
}
}
# Agent discovers and uses MCP tools automatically
agent = project.create_agent(
name="data-analyst",
model="gpt-4",
tool_sources=[mcp_config]
)Claude Desktop and Claude Code Integration
Claude Desktop provides the simplest MCP integration path through configuration files. Add your server to claude_desktop_config.json, restart Claude, and your tools become available immediately. Claude Code extends this to development environments with IDE-integrated MCP support.
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
{
"mcpServers": {
"production-server": {
"type": "http",
"url": "https://your-mcp-server.com/mcp",
"headers": {
"Authorization": "Bearer ${MCP_API_KEY}"
}
},
"local-dev-server": {
"command": "node",
"args": ["/path/to/your/server/index.js"]
}
}
}Cross-Platform Tool Orchestration
MCP’s protocol-based design enables sophisticated cross-platform scenarios. An agent in Azure Foundry can invoke tools from MCP servers while a Claude Desktop user accesses the same servers simultaneously. The protocol ensures consistent behavior regardless of client platform.
graph TB
subgraph "AI Platforms"
AF[Azure AI Foundry]
CD[Claude Desktop]
CC[Claude Code]
end
subgraph "MCP Protocol Layer"
MP[MCP Protocol]
end
subgraph "Your MCP Servers"
S1[Database Server]
S2[API Server]
S3[Analytics Server]
end
AF --> MP
CD --> MP
CC --> MP
MP --> S1
MP --> S2
MP --> S3Best Practices for Platform Integration
Design MCP servers for multi-platform use by implementing standard authentication, providing clear tool descriptions, handling errors gracefully, logging all platform interactions, and testing across all target platforms. Monitor usage patterns across platforms to optimize performance and identify integration issues early.
Conclusion
Integrating MCP servers with Azure AI Foundry and Claude enables powerful AI workflows backed by your custom data and tools. The protocol’s platform independence ensures your investment in MCP servers delivers value across the entire AI ecosystem. The final article in this series covers production deployment strategies and operational best practices for maintaining MCP servers at scale.
