Getting started

Installation

Prerequisites: Python 3.10 or later, uv, Access to a Valkey datastore

Configure your MCP client: Add the following to your MCP client configuration:

{
  "mcpServers": {
    "valkey": {
      "command": "uvx",
      "args": ["awslabs.valkey-mcp-server@latest"],
      "env": {
        "VALKEY_HOST": "<your-endpoint>",
        "VALKEY_USE_SSL": "true"
      }
    }
  }
}

Replace <your-endpoint> with the endpoint of your Valkey datastore. Your coding agent must have network access to the endpoint.

Connecting to Amazon ElastiCache: If your Valkey serverless cache has a public endpoint, use the public endpoint address as VALKEY_HOST. No tunnel is required. The connection requires TLS 1.3 and an IAM authentication token. For details on generating the token, see Connect to a cache with a public endpoint.

For all other ElastiCache for Valkey caches, you cannot reach the cache directly from your local machine. For SSM tunnel setup, IAM auth token generation, and TLS handling in tunnel mode, see the ElastiCache connection guide in the AWS Agent Toolkit, or install the Amazon ElastiCache Skill to let your agent handle the connection setup end-to-end. After the tunnel is established, use VALKEY_HOST=127.0.0.1 in the config above.

Restart your MCP client after saving the configuration.

You can prompt your agent to verify the connection with:

Use the Valkey MCP server to verify the connection to my datastore and confirm that Valkey operations are available.

Configuration

Configure the Valkey MCP server through environment variables in your MCP client configuration. You can also use the --readonly flag to restrict agent access to read-only operations.