Skip to main content
This guide will walk you through setting up your first MCP server and tool in AgentPass Console. You’ll be able to connect an AI agent to your MCP server in just a few steps.
We’ll use OpenWeatherMap APIs as an example in this guide for ease of use. You can use your own APIs or services.

Before you start

  • Estimated time: 15 minutes
  • If you don’t have an AgentPass account, you can sign up here.
  • If you don’t have an API key from OpenWeatherMap, you can sign up for a free account here. After signing up, you can get your API key from the API Keys page.

Step 1: Access the Console

1

Login to AgentPass Console

Navigate to console and sign in with your credentials.

Step 2: Create Your First MCP Server

1

Click Create Server

From the Servers page, click the “Create Server” button to start the server creation process.
2

Configure Server Details

Fill in the basic server information:
  • Name: My First MCP Server
  • Description: A server for testing MCP integration
3

Create the Server

Click “Create Server” to finalize your MCP server setup.

Step 3: Create Your First Tool

1

Create New Tool

Now you’ve landed on the Flow Builder. Here you can create your first tool.
  • Click “Create your first tool”
  • Choose “Create manually” for full control
2

Configure Tool Basics

Enter the tool information:
  • Name: get_weather
  • Description: Get current weather for a location
  • API URL: https://api.openweathermap.org/data/2.5/weather
  • HTTP Method: GET
  • Click “Create Tool”
Create new tool form showing basic configuration options
3

Define Parameters

In the flow builder, you can see the tool you just created.
  • Hover over the tool and click “Edit Tool” (pencil icon)
  • Go to the Tab “Tool Parameters”
Add the following parameter:
  • Name: q
  • Data type: string
  • Description: Location
  • Click “Add Parameter” and then click “Save Changes”
Tool parameter configuration
4

Configure HTTP Request

Move to the “HTTP Request” tab. This is where the API request will be configured.Add a new parameter (Tab Params):
  • Name: appid
  • Value: your-api-key (remember to replace with your actual API key)
  • Click “Save Changes”
HTTP Request configuration

Step 4: Test Your Tool

1

Access Tool Playground

From the flow builder, hover your mcp tool and click the “Test” button to open the playground.
2

Enter Test Parameters

In the playground interface:
  • Enter a test value for the location parameter (e.g., “London”)
3

Execute Test

Click “Execute” to test your tool against the actual API.
4

Review Results

Examine the response to ensure your tool is working correctly:
  • Check the HTTP status code
  • Review the response data
Tool test results
Congratulations! You’ve successfully created and tested your first MCP tool.

Bonus Step: Connect an AI Agent

1

Choose Your AI Tool

In the flow builder, in the AI Agent node,click “Connect”
2

Follow Integration Guide

Follow the specific integration guide for your chosen AI tool to add the MCP server connection.
3

Test the Integration

In your AI tool, try asking it to use your weather tool:
"Can you get the current weather for San Francisco?"

Need Help?

  • Visit our concepts guide to understand MCP fundamentals
  • Check out specific feature guides in the documentation sidebar
  • Contact support@agentpass.ai for technical assistance
You’re now ready to build more sophisticated MCP integrations with AgentPass Console!
I