IntegrationsMCP Integration

MCP Integration

Connect Model Context Protocol (MCP) servers to extend your agent with custom tools and capabilities.

Introduction

The MCP (Model Context Protocol) integration enables you to connect MCP servers to your LiveTok AI Agent, extending its capabilities with custom tools and functions. This allows you to integrate specialized systems, proprietary data, and custom workflows that aren’t available through standard integrations.

Purpose: Provide a flexible framework for extending AI capabilities beyond standard integrations, enabling custom tools, specialized knowledge bases, and advanced automation.

Motivation: Every business has unique systems, data sources, and workflows. The MCP integration provides a standardized way to connect these custom resources to your AI Agent, enabling highly customized conversational experiences.


Capabilities

Custom Tools and Functions

Connect MCP servers that provide custom tools and functions your AI Agent can use during conversations. These tools can integrate with proprietary systems, internal APIs, and specialized business logic.

Extended Capabilities

Extend your AI Agent’s capabilities beyond standard integrations by connecting MCP servers that provide domain-specific tools, data access, and functionality tailored to your business needs.

Flexible Integration

Connect any system or service that implements the Model Context Protocol, enabling integration with custom-built tools, specialized databases, and proprietary systems.


Configuration

Prerequisites

  • MCP server URL
  • Understanding of available tools from your MCP server
  • Access to the MCP server endpoint

Connect MCP

  1. Set Up MCP Server

    • Ensure you have an MCP server running and accessible
    • Note the server URL (e.g., http://your-server.com/mcp or ws://your-server.com/mcp)
  2. Get Available Tools

    • Connect to your MCP server to discover available tools
    • Note the tool names and descriptions you want to use
  3. Configure in LiveTok

    • Go to Dashboard > Connect
    • Find MCP Integration in the available tools
    • Click Configure
    • Enter your Server URL
    • Select the Tools you want to enable from the available list
    • Click Save
  4. Verify Connection

    • Check that MCP Integration is listed under “Configured Tools”
    • Test by using one of the configured tools in a conversation

Configuration Settings

  • Server URL: The URL of your MCP server endpoint
  • Selected Tools: The specific tools from your MCP server that you want to enable for your AI Agent

Once configured, your AI Agent can use the selected MCP tools during conversations to extend its capabilities.