Kong Konnect MCP Server

What does “Kong Konnect” MCP Server do?

The Kong Konnect MCP Server is a Model Context Protocol (MCP) server that enables AI assistants, such as Claude, to interact directly with Kong Konnect’s API Gateway. This server acts as a bridge, empowering AI-driven tools to query and analyze gateway configurations, traffic, and analytics through natural language. Developers can use the server to retrieve analytics data, inspect API gateway configurations, and manage control planes, all via standardized MCP tools. By integrating external data sources like the Kong Konnect APIs into AI workflows, this MCP server streamlines tasks such as monitoring traffic, auditing configurations, and managing API services, enhancing the efficiency and intelligence of development and operations work.

List of Prompts

• No explicit prompt templates are documented in the repository files or README.

List of Resources

• No explicit MCP resources are documented in the repository files or README.

Analytics Tools

• Query API Requests
  Query and analyze Kong API Gateway requests with customizable filters such as time range, status codes, HTTP methods, consumer IDs, and more.

• Get Consumer Requests
  Analyze API requests made by a specific consumer, filtered by consumer ID and time range.

Configuration Tools

• (Other configuration tools are referenced but not detailed in the README or project files.)

Control Planes Tools

• (Tools related to control plane management are referenced but not detailed in the README or project files.)

Use Cases of this MCP Server

• API Analytics and Reporting
  Developers and operations teams can use the server to retrieve detailed analytics on API requests, including filtering by status codes, methods, consumers, services, and more for comprehensive monitoring.

• Consumer Behavior Analysis
  Analyze requests and behaviors for individual consumers, helping support, billing, or security audit tasks.

• Gateway Configuration Auditing
  List and inspect services, routes, consumers, and plugins to audit and verify API gateway configurations.

• Control Plane Management
  Manage and inspect control planes and control plane groups, simplifying the administration of distributed API gateways.

• AI-Driven Operations
  Enable AI assistants to perform real-time queries and diagnostics on API infrastructure, reducing manual effort and accelerating incident response.

How to set it up

Windsurf

1. Prerequisites: Ensure Node.js 20+ and MCP-compatible client are installed.
2. Clone & Install:

   git clone https://github.com/Kong/mcp-konnect.git
   cd mcp-konnect
   npm install
   npm run build
   

3. Set Environment Variables:

   export KONNECT_ACCESS_TOKEN=your_api_key
   export KONNECT_REGION=us
   

4. Add MCP Server in Windsurf config:
   Edit your Windsurf configuration file (e.g., windsurf.json):

   {
     "mcpServers": {
       "kong-konnect": {
         "command": "npx",
         "args": ["@kong/mcp-konnect@latest"],
         "env": {
           "KONNECT_ACCESS_TOKEN": "${KONNECT_ACCESS_TOKEN}",
           "KONNECT_REGION": "us"
         }
       }
     }
   }
   

5. Restart Windsurf and verify.

Claude

1. Prerequisites: Node.js 20+, Claude Desktop or compatible client.
2. Clone, Install, and Build: (see above)
3. Set Env Vars: (see above)
4. Edit Claude config:

   {
     "mcpServers": {
       "kong-konnect": {
         "command": "npx",
         "args": ["@kong/mcp-konnect@latest"],
         "env": {
           "KONNECT_ACCESS_TOKEN": "${KONNECT_ACCESS_TOKEN}",
           "KONNECT_REGION": "us"
         }
       }
     }
   }
   

5. Restart Claude and verify.

Cursor

1. Prerequisites: Node.js 20+, Cursor with MCP capability.
2. Clone, Install, Build: (see above)
3. Set Env Vars: (see above)
4. Configure in Cursor:

   {
     "mcpServers": {
       "kong-konnect": {
         "command": "npx",
         "args": ["@kong/mcp-konnect@latest"],
         "env": {
           "KONNECT_ACCESS_TOKEN": "${KONNECT_ACCESS_TOKEN}",
           "KONNECT_REGION": "us"
         }
       }
     }
   }
   

5. Restart Cursor and verify.

Cline

1. Prerequisites: Node.js 20+, Cline with MCP support.
2. Clone, Install, Build: (see above)
3. Set Env Vars: (see above)
4. Add to Cline config:

   {
     "mcpServers": {
       "kong-konnect": {
         "command": "npx",
         "args": ["@kong/mcp-konnect@latest"],
         "env": {
           "KONNECT_ACCESS_TOKEN": "${KONNECT_ACCESS_TOKEN}",
           "KONNECT_REGION": "us"
         }
       }
     }
   }
   

5. Restart and test connection.

Note:
Always secure API keys by using environment variables. Example:

{
  "mcpServers": {
    "kong-konnect": {
      "command": "npx",
      "args": ["@kong/mcp-konnect@latest"],
      "env": {
        "KONNECT_ACCESS_TOKEN": "${KONNECT_ACCESS_TOKEN}",
        "KONNECT_REGION": "us"
      },
      "inputs": {
        "KONNECT_ACCESS_TOKEN": "env:KONNECT_ACCESS_TOKEN"
      }
    }
  }
}

How to use this MCP inside flows

Using MCP in FlowHunt

To integrate MCP servers into your FlowHunt workflow, start by adding the MCP component to your flow and connecting it to your AI agent:

Click on the MCP component to open the configuration panel. In the system MCP configuration section, insert your MCP server details using this JSON format:

{
  "kong-konnect": {
    "transport": "streamable_http",
    "url": "https://yourmcpserver.example/pathtothemcp/url"
  }
}

Once configured, the AI agent is now able to use this MCP as a tool with access to all its functions and capabilities. Remember to change “kong-konnect” to whatever the actual name of your MCP server is and replace the URL with your own MCP server URL.

Overview

Based on the above, Kong Konnect MCP Server is a work-in-progress but already exposes relevant tools for analytics and configuration. However, resources and prompts are not documented, and support for roots or sampling is not mentioned. The project is open source, reasonably active, and has clear setup instructions.

Our opinion

Kong Konnect MCP Server integrates well with AI workflows for API operations and analytics. However, the lack of documentation for prompts and resources, as well as no explicit mention of roots or sampling, limits its completeness. Still, it’s a functional and valuable MCP for its domain.

Rating: 6/10

MCP Score