mcp-confluent

An MCP server implementation that enables AI assistants to interact with Confluent Kafka and Confluent Cloud REST APIs. This server allows AI tools like Claude Desktop and Goose CLI to manage Kafka topics, connectors, and Flink SQL statements through natural language interactions.

Demo

Goose CLI

Claude Desktop

Table of Contents

• mcp-confluent
  • Demo
    • Goose CLI
    • Claude Desktop
  • Table of Contents
  • User Guide
    • Getting Started
    • Configuration
    • Environment Variables Reference
    • Usage
    • Configuring Claude Desktop
    • Configuring Goose CLI
  • Developer Guide
    • Project Structure
    • Building and Running
    • Testing
      • MCP Inspector
    • Adding a New Tool
    • Generating Types
    • Contributing

User Guide

Getting Started

1. Create a .env file: Copy the example .env file structure (shown below) into a new file named .env in the root of your project.

2. Populate the .env file: Fill in the necessary values for your Confluent Cloud environment. See the Configuration section for details on each variable.

3. Install Node.js (if not already installed)

   • We recommend using NVM (Node Version Manager) to manage Node.js versions
   • Install and use Node.js:

   nvm install 22
   nvm use 22
   

Configuration

Create a .env file in the root directory of your project with the following configuration:

# .env file
BOOTSTRAP_SERVERS="pkc-v12gj.us-east4.gcp.confluent.cloud:9092"
KAFKA_API_KEY="..."
KAFKA_API_SECRET="..."
KAFKA_REST_ENDPOINT="https://pkc-v12gj.us-east4.gcp.confluent.cloud:443"
KAFKA_CLUSTER_ID=""
KAFKA_ENV_ID="env-..."
FLINK_ENV_ID="env-..."
FLINK_ORG_ID=""
FLINK_REST_ENDPOINT="https://flink.us-east4.gcp.confluent.cloud"
FLINK_ENV_NAME=""
FLINK_DATABASE_NAME=""
FLINK_API_KEY=""
FLINK_API_SECRET=""
FLINK_COMPUTE_POOL_ID="lfcp-..."
CONFLUENT_CLOUD_API_KEY=""
CONFLUENT_CLOUD_API_SECRET=""
CONFLUENT_CLOUD_REST_ENDPOINT="https://api.confluent.cloud"
SCHEMA_REGISTRY_API_KEY="..."
SCHEMA_REGISTRY_API_SECRET="..."
SCHEMA_REGISTRY_ENDPOINT="https://psrc-zv01y.northamerica-northeast2.gcp.confluent.cloud"

Environment Variables Reference

Usage

This MCP server is designed to be used with various MCP clients, such as Claude Desktop or Goose CLI/Desktop. The specific configuration and interaction will depend on the client you are using. However, the general steps are:

1. Build: Follow the instructions in the Developer Guide to build and run the server from source. This typically involves:

   • Installing dependencies (npm install)
   • Building the project (npm run build or npm run dev)

2. Configure your MCP Client: Each client will have its own way of specifying the MCP server's address and any required credentials. You'll need to configure your client (e.g., Claude, Goose) to connect to the address where this server is running (likely localhost with a specific port). The port the server runs on may be configured by an environment variable.

3. Start the MCP Client: Once your client is configured to connect to the MCP server, you can start your mcp client and on startup - it will stand up an instance of this MCP server locally. This instance will be responsible for managing data schemas and interacting with Confluent Cloud on your behalf.

4. Interact with Confluent through the Client: Once the client is connected, you can use the client's interface to interact with Confluent Cloud resources. The client will send requests to this MCP server, which will then interact with Confluent Cloud on your behalf.

Configuring Claude Desktop

See here for more details about installing Claude Desktop and MCP servers.

To configure Claude Desktop to use this MCP server:

1. Open Claude Desktop Configuration

   • On Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
   • On Windows: %APPDATA%\Claude\claude_desktop_config.json

2. Edit Configuration File

   • Open the config file in your preferred text editor
   • Add or modify the configuration using one of the following methods:
   Option 1: Run from source

   {
     "mcpServers": {
       "confluent": {
         "command": "node",
         "args": [
           "/path/to/confluent-mcp-server/dist/index.js",
            "--env-file",
           "/path/to/confluent-mcp-server/.env",
         ]
       }
     }
   }
   

   Option 2: Run from npx

   {
     "mcpServers": {
       "confluent": {
         "command": "npx",
         "args": [
           "-y"
           "@confluentinc/mcp-confluent",
           "-e",
           "/path/to/confluent-mcp-server/.env"
         ]
       }
     }
   }
   

   Replace /path/to/confluent-mcp-server/ with the actual path where you've installed this MCP server.

3. Restart Claude Desktop

   • Close and reopen Claude Desktop for the changes to take effect
   • The MCP server will automatically start when Claude Desktop launches

Now Claude Desktop will be configured to use your local MCP server for Confluent interactions.

Configuring Goose CLI

See here for detailed instructions on how to install the Goose CLI.

Once installed, follow these steps:

1. Run the Configuration Command:

   goose configure
   

2. Follow the Interactive Prompts:

   • Select Add extension
   • Choose Command-line Extension
   • Enter mcp-confluent as the extension name
   • Choose one of the following configuration methods:
   Option 1: Run from source

   node /path/to/confluent-mcp-server/dist/index.js --env-file /path/to/confluent-mcp-server/.env
   

   Option 2: Run from npx

   npx -y @confluentinc/mcp-confluent -e /path/to/confluent-mcp-server/.env
   

Replace /path/to/confluent-mcp-server/ with the actual path where you've installed this MCP server.

Developer Guide

Project Structure

/
├── src/                 # Source code
│   ├── confluent/       # Code related to Confluent integration (API clients, etc.)
│   ├── tools/           # Tool implementations (each tool in a separate file)
│   ├── index.ts         # Main entry point for the server
│   └── ...              # Other server logic, utilities, etc.
├── dist/                # Compiled output (TypeScript -> JavaScript)
├── openapi.json         # OpenAPI specification for Confluent Cloud
├── .env                 # Environment variables (example - should be copied and filled)
├── README.md            # This file
└── package.json         # Node.js project metadata and dependencies

Building and Running

1. Install Dependencies:

   npm install
   

2. Development Mode (watch for changes):

   npm run dev
   

   This command compiles the TypeScript code to JavaScript and automatically rebuilds when changes are detected in the src/ directory.

3. Production Build (one-time compilation):

   npm run build
   

4. Start the Server:

   npm run start
   

Testing

MCP Inspector

For testing MCP servers, you can use MCP Inspector which is an interactive developer tool for testing and debugging MCP servers.

# make sure you've already built the project either in dev mode or by running npm run build
npx @modelcontextprotocol/inspector node  $PATH_TO_PROJECT/dist/index.js --env-file $PATH_TO_PROJECT/.env

Adding a New Tool

1. Add a new enum to the enum class ToolName.
2. Add your new tool to the handlers map in the ToolFactory class.
3. Create a new file, exporting the class that extends BaseToolHandler.
   1. Implement the handle method of the base class.
   2. Implement the getToolConfig method of the base class.
4. Once satisfied, add it to the set of enabledTools in index.ts.

Generating Types

# as of v7.5.2 there is a bug when using allOf w/ required https://github.com/openapi-ts/openapi-typescript/issues/1474. need --empty-objects-unknown flag to avoid it
npx openapi-typescript ./openapi.json -o ./src/confluent/openapi-schema.d.ts --empty-objects-unknown

Contributing

Bug reports and feedback is appreciated in the form of Github Issues. For guidelines on contributing please see CONTRIBUTING.md