Skip to main content
🎆

See the MCP Server in action with our live demo - no setup needed!

Prerequisites

  1. Alloy Account: Sign in
  2. MCP Client: Claude Desktop (Download here), your backend application, or any MCP-compatible client
  3. Credentials to Connectors: Ensure you have access to creating credentials for the connectors (Slack, HubSpot, Notion, etc.) you’re looking to support or interact with.

Step 1: Get Your API Key (1 minute)

1
2

Navigate to Settings → API Keys
3

Click “Generate API Key”
4

Copy and save your API Key

Step 2: Create Your First Server (1 minute)

Open your terminal and run this command (replace YOUR_API_KEY with your API key):
curl -X POST https://mcp.runalloy.com/api/servers \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "My First MCP Server",
    "description": "Connects AI assistants to business platforms"
  }'
You should see:
{
  "serverId": "my-first-mcp-server-a1b2c3d4",
  "name": "My First MCP Server",
  "description": "Connects AI assistants to business platforms",
  "accessToken": "mcp_token_xyz789",
  "url": "https://mcp.runalloy.com/mcp/my-first-mcp-server-a1b2c3d4",
  "accessUrl": "https://mcp.runalloy.com/mcp/my-first-mcp-server-a1b2c3d4/mcp_token_xyz789"
}
Important: Be sure to copy the accessUrl field as this is your server’s unique address.

Step 3: Connect Your MCP Client (1 minute)

For Claude Desktop:

1

Open Claude Desktop
2

Go to Settings → Connectors → Add Custom Connector
3

Paste your accessUrl in the “Remote MCP server URL” input box and name the connector.
4

Click “Add”
5

Your Alloy MCP server should now be accessible and you can click “Configure” to manage tools.

For Cursor (paid plan required):

Add to your .cursorrules or workspace settings:
{
  "mcp": {
    "servers": {
      "alloy": {
        "url": "YOUR_ACCESS_URL_HERE"
      }
    }
  }
}

For Python Applications:

import requests
import json

# Your server details from Step 2
ACCESS_URL = "https://mcp.runalloy.com/mcp/my-first-mcp-server-a1b2c3d4/mcp_token_xyz789"

# List available tools
response = requests.post(ACCESS_URL, json={
    "jsonrpc": "2.0",
    "method": "tools/list",
    "id": 1
})

tools = response.json()
print(f"Available tools: {json.dumps(tools, indent=2)}")

# Execute an action (example: list Slack channels)
response = requests.post(ACCESS_URL, json={
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
        "name": "execute_action_alloy",
        "arguments": {
            "connectorId": "slack",
            "actionId": "conversations_list",
            "parameters": {"limit": 10}
        }
    },
    "id": 2
})

result = response.json()
print(f"Slack channels: {json.dumps(result, indent=2)}")

Step 4: Connect to Your Platform (1 minute)

To connect your MCP server to a platform, in your MCP client, type:
"Show me available platforms to connect"
The AI assistant will list available integrations. You can then choose to connect to a specific platform. The assistant will guide you through the authentication process:
  • For OAuth platforms: You’ll receive an authorization link
  • For API key platforms: You’ll be prompted to provide credentials
Once connected, the model should confirm that the connection was successful: “I’ve connected [Platform]“

Step 5: Your First Commands (1 minute)

Try these commands with your AI assistant:

Example Actions

For Slack:
"Send 'Hello from the MCP Server!' to the #general channel in Slack"
For HubSpot:
"Create a new contact in HubSpot with email john@example.com"
For Notion:
"Create a new page in Notion titled 'Meeting Notes'"
Each platform has different capabilities. Ask your AI assistant:
"What actions can I perform with [Platform]?"

Congratulations!

You’ve just:
  • Created an MCP server
  • Connected an AI assistant to a business platform
  • Performed automated actions through natural language

What’s Next?

Try More Commands

Get Creative:
"Create a task in [Your Platform]"
"Update a record with new information"
"Send a notification to your team"

Add More Platforms

"Show me available integrations"
"Help me connect to HubSpot"
"What can you do with Notion?"

Explore Advanced Features

Set Restrictions - Limit which platforms the AI can access:
curl -X PUT https://mcp.runalloy.com/api/servers/YOUR_SERVER_ID \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "restrictions": {
      "connectors": {
        "mode": "allowlist",
        "allowedIds": ["slack", "notion"]
      }
    }
  }'
I