azure-logic-apps-mcp by iLoveAgents - MCP Server

Azure Logic Apps MCP Server

Turn Azure Logic Apps workflows into tools that AI assistants can use through the Model Context Protocol (MCP).

Let AI Agents Do the Work

Don't read this whole README! Just download and give it to an AI agent.

Then say: "Set up my Azure Logic App as an MCP server"

The agent will handle everything:

Ask for your Azure resources
Configure MCP endpoints
Set up authentication (dev or production)
Create your first tool
Connect with VS Code or Claude

No dependencies. No scripts. Just AGENTS.md and an AI agent.

Still want to do it manually? Continue reading...

Manual Setup (If You Prefer)

Prerequisites

Azure subscription
Azure Logic App (Standard) - Create one
Azure CLI - Install guide

1. Enable MCP on Your Logic App

# Set your values
export SUBSCRIPTION_ID="your-subscription-id"
export RESOURCE_GROUP="your-resource-group"
export LOGIC_APP_NAME="your-logic-app-name"

# Login and get token
az login
TOKEN=$(az account get-access-token --resource https://management.azure.com --query accessToken -o tsv)

# Get Kudu URL
KUDU_URL=$(az logicapp show \
  --name $LOGIC_APP_NAME \
  --resource-group $RESOURCE_GROUP \
  --query "enabledHostNames[?contains(@, 'scm')]|[0]" \
  -o tsv)

# Create host.json
cat > host.json <<'EOF'
{
  "version": "2.0",
  "extensionBundle": {
    "id": "Microsoft.Azure.Functions.ExtensionBundle.Workflows",
    "version": "[1.*, 2.0.0)"
  },
  "extensions": {
    "workflow": {
      "McpServerEndpoints": {
        "enable": true,
        "authentication": {
          "type": "anonymous"
        }
      }
    }
  }
}
EOF

# Upload and restart
curl -X PUT "https://${KUDU_URL}/api/vfs/site/wwwroot/host.json" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -H "If-Match: *" \
  --data-binary @host.json

az logicapp restart --name $LOGIC_APP_NAME --resource-group $RESOURCE_GROUP

Wait 30 seconds for the restart.

2. Create Your First Tool

See for the complete workflow template and upload instructions.

3. Connect to AI Assistants

VS Code (Recommended):

Command Palette → "MCP: Add Server"
Enter: https://your-logic-app-name.azurewebsites.net/api/mcp
Your tools appear automatically

Claude Desktop:

{
  "mcpServers": {
    "logic-apps": {
      "url": "https://your-logic-app-name.azurewebsites.net/api/mcp",
      "transport": "streamable-http"
    }
  }
}

MCP Inspector:

npx @modelcontextprotocol/inspector
# URL: https://your-logic-app-name.azurewebsites.net/api/mcp
# Transport: Streamable HTTP

Creating Tools

Each tool is a stateless or Stateful workflow with:

HTTP Request Trigger (named "manual") with JSON schema
Business Logic (any workflow actions)
HTTP Response (200 status, JSON body)

For complete workflow templates and examples, see .

Use Cases

Data Integration: Query databases, APIs, Azure services
Business Automation: Trigger workflows, send notifications
Data Transformation: Convert formats, process data
Custom Actions: Execute specialized business logic

Production Security

The quick start uses anonymous authentication for development/testing.

For production, configure OAuth 2.0 authentication:

Industry-standard security
Azure AD integration
User-based access control
Scope-based permissions

See for complete OAuth 2.0 setup instructions with:

Azure AD app registration
Easy Auth configuration
VS Code auto-discovery support
Troubleshooting guide

Note: VS Code 1.102+ auto-discovers OAuth settings - just add the server URL and sign in!

Troubleshooting

Tool not appearing:

Verify workflow is in <tool-name>/workflow.json
Check "kind": "Stateless" is set
Restart Logic App and wait 30 seconds

Authentication errors:

Verify authentication is inside McpServerEndpoints in host.json
Restart Logic App after configuration changes

For detailed troubleshooting and automated deployment, see

Resources

License

MIT