Introduction to MCP & the Alloy MCP Server

​

What is MCP?

MCP (Model Context Protocol) is a standardized protocol developed by Anthropic that allows AI assistants to: Think of MCP as a universal adapter that lets AI assistants interact with any system that speaks the protocol. You can read more about MCP here.

​

How Alloy Implements MCP

Alloy’s MCP Server acts as a bridge between MCP clients and the integrations in the Alloy platform:

​

Key Features

​

🚀 Instant Access to 250+ Integrations

Connect Claude to Slack, HubSpot, Salesforce, Notion, and hundreds more platforms without writing any code.

​

🔐 Enterprise-Grade Security

• OAuth 2.0 authentication for secure platform connections
• AES-256 encryption for stored credentials
• Per-request isolation to prevent data leakage
• Fine-grained access controls and restrictions

​

🎯 Smart Tool Discovery

MCP clients can automatically discover available actions and learn how to use them through the MCP protocol.

​

👤 Multi-User Support

Each user gets their own isolated data context, perfect for teams and organizations.

​

⚡ Real-Time Operations

Execute actions instantly - send messages, create records, fetch data - all in real-time.

​

Core Concepts

​

1. Servers

An MCP server is an instance that MCP clients connect to. Each server can have:

• Access restrictions to control what the AI can do
• Unique authentication tokens for security
• Public or private access modes

​

2. Tools

Tools are the actions MCP clients can perform. In connectivity mode, you get 8 powerful tools:

• list_connectors_alloy - Discover available platforms and integrations
• get_connector_resources_alloy - See what actions are available for each platform
• get_action_details_alloy - Get detailed parameter information for actions
• execute_action_alloy - Perform actions on connected platforms
• create_credential_alloy - Set up new platform connections via OAuth or API keys
• get_credentials_alloy - List existing credentials for a connector
• get_credential_metadata_alloy - Get authentication requirements for platforms
• discover_action_path_alloy - Get intelligent workflow suggestions

​

3. Connectors

Connectors are the 250+ pre-built integrations with platforms like:

• Communication: Slack, Microsoft Teams, Twilio,
• CRM: Salesforce, HubSpot, ServiceNow
• E-commerce: Shopify, WooCommerce, BigCommerce
• Productivity: Notion, JIRA, ClickUp
• And many more…

​

4. Restrictions

Control what MCP clients can access with three types of restrictions:

• Connector restrictions: Limit which platforms the AI can use
• Action restrictions: Block specific actions (like deleting data)
• User restrictions: Control access for specific users

​

Common Use Cases

​

Customer Support Automation

Let AI assistants handle support tickets by connecting to Zendesk, searching knowledge bases, and drafting responses.

​

Sales Intelligence

Connect AI assistants to your CRM to get instant insights, update deals, and track customer interactions.

​

Content Management

Use AI assistants to manage content across Notion, WordPress, and other platforms simultaneously.

​

Workflow Automation

Create complex workflows where AI assistants orchestrate actions across multiple platforms.

​

Data Analysis

Pull data from various sources and have AI assistants analyze and report on trends.

​

Getting Started

Ready to connect AI assistants to your business apps? Here’s how to begin:

​

Why Choose Alloy MCP?

• No Code Required - Connect platforms without writing integration code
• Production Ready - Built on Alloy’s battle-tested infrastructure
• Instant Updates - New connectors added regularly
• Enterprise Support - Get help when you need it
• SOC 2 Compliant - Enterprise-grade security and compliance

​

Next Steps

The Alloy MCP Server is built and maintained by Alloy Automation, the leading embedded integration platform.