home / mcp / kong konnect mcp server
A Model Context Protocol (MCP) server for interacting with Kong Konnect APIs, allowing AI assistants to query and analyze Kong Gateway configurations, traffic, and analytics.
Configuration
View docs{
"mcpServers": {
"kong-mcp-konnect": {
"command": "node",
"args": [
"/absolute/path/to/mcp-konnect/build/index.js"
],
"env": {
"KONNECT_REGION": "us",
"KONNECT_ACCESS_TOKEN": "kpat_api_key_here"
}
}
}
}You run a Kong Konnect MCP Server to let AI agents query and analyze Kong Gateway configurations, traffic, and analytics through natural language. It exposes a set of practical tools to inspect analytics, configurations, and control planes, enabling you to build interactive workflows with MCP-compatible assistants.
Start by running the MCP server locally and connecting your MCP client (such as Claude Desktop) to it. You will expose a stdio-based MCP entry that the client can launch and communicate with through a shared process. Use the provided tools to query analytics, list services, routes, consumers, plugins, and manage control planes and their groups. You can weave these tools into natural language prompts to explore configurations and performance data across your Kong Konnect environment.
Prerequisites: Node.js 20 or higher is required to build and run the MCP server.
Step 1: Install Node.js 20+ on your machine or in your development environment.
Step 2: Prepare your project directory and install dependencies.
Step 3: Build and run the MCP server.
To connect Claude Desktop to this MCP server, configure Claude to launch the local MCP process as shown in the example configuration.
1) Ensure Claude is installed for Desktop.
2) Create or edit Claude Desktop configuration with a stdio MCP entry that points to your local build index. The command should start the MCP server process and provide the necessary environment variables.
Analyze API traffic by querying the analytics tools for a time range and filters, then drill down into specific services, routes, or consumers to understand error patterns and usage.
Your MCP server can be configured to pass the required API key and region to the Kong Konnect API. Use environment variables to securely provide credentials at runtime.
If you encounter connection or authentication issues, verify that your API key is valid, the region is correct, and the MCP process has access to the environment variables. Check control plane IDs and time ranges for analytics queries to ensure resources exist and data is available.
Development notes describe how to extend tools, add new parameters, and implement new operations. Define parameters, document tools, implement logic in the operations directory, and register the tool in the central tool registry.
This MCP server is designed to enable interactive AI-assisted querying of Kong Konnect configurations, analytics, and control planes. It leverages a stdio-based entry point that your MCP client can launch and manage alongside your workflow tooling.
Query API request analytics with filters such as time range, status codes, HTTP methods, consumer/service/route IDs, and limits to results.
Analyze API requests made by a specific consumer over a chosen time range with options to filter by success or failure and limit results.
List all services under a specific control plane with pagination controls.
List all routes under a specific control plane with pagination controls.
List all consumers under a specific control plane with pagination controls.
List all plugins under a specific control plane with pagination controls.
List all control planes in your organization with filtering, sorting, and pagination.
Get detailed information about a specific control plane.
List members of a specific control plane group with pagination.
Check if a control plane is a member of any group.