Skip to main content

​
Custom MCP Integration

Connect your own MCP (Model Context Protocol) servers to extend the AI agent with custom tools.

​
What is MCP?

MCP (Model Context Protocol) is a standard protocol for connecting AI agents to external tools and data sources. AccelOS uses MCP to communicate with all integrations.

​
Use Cases

  • Connect internal APIs and services
  • Add custom data sources
  • Integrate proprietary tools
  • Extend agent capabilities with custom logic

​
Setup

1

Deploy Your MCP Server

Deploy an MCP server that implements the tools you want to expose.
2

Add Integration in AccelOS

  1. Go to Integrations → Custom MCP
  2. Enter a name for your integration
  3. Enter the MCP server URL
  4. Configure authentication if required
3

Test and Save

Click Test Connection to verify, then Save.

​
MCP Server Requirements

Your MCP server should:
  • Implement the MCP protocol specification
  • Be accessible from AccelOS (check network/firewall rules)
  • Return tool definitions and handle tool calls
  • Support appropriate authentication

​
Authentication

Custom MCP integrations support:
  • API Key: Pass a static API key
  • Bearer Token: Use OAuth2 tokens
  • Basic Auth: Username and password

​
Example

A custom MCP server might expose tools like: 
{
  "tools": [
    {
      "name": "query_internal_metrics",
      "description": "Query internal metrics database",
      "parameters": {
        "metric_name": "string",
        "time_range": "string"
      }
    },
    {
      "name": "check_feature_flags",
      "description": "Check feature flag status",
      "parameters": {
        "flag_name": "string"
      }
    }
  ]
}
Contact your AccelOS representative for guidance on building custom MCP integrations.