stdio databaseawsai

AWS KB Retrieval MCP Server

Query Amazon Bedrock Knowledge Bases for RAG-style document retrieval.

claude_desktop_config.json

{
  "mcpServers": {
    "server-aws-kb-retrieval": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-aws-kb-retrieval"
      ],
      "env": {
        "AWS_ACCESS_KEY_ID": "your-access-key-id",
        "AWS_SECRET_ACCESS_KEY": "your-secret-access-key",
        "AWS_REGION": "us-east-1",
        "KNOWLEDGE_BASE_ID": "your-kb-id"
      }
    }
  }
}

Environment Variables

Variable Required Description
AWS_ACCESS_KEY_ID ✓ Required AWS access key ID with Bedrock permissions.
AWS_SECRET_ACCESS_KEY ✓ Required AWS secret access key paired with the access key ID.
AWS_REGION ✓ Required AWS region where the Knowledge Base is deployed.
KNOWLEDGE_BASE_ID ✓ Required The Amazon Bedrock Knowledge Base ID to query.

Setup Instructions

  1. Ensure Node.js 18+ is installed on your system.
  2. Copy the config block above into your claude_desktop_config.json file (typically at ~/.config/claude/claude_desktop_config.json on macOS/Linux or %APPDATA%\Claude\claude_desktop_config.json on Windows).
  3. Set the AWS_ACCESS_KEY_ID environment variable: AWS access key ID with Bedrock permissions.
  4. Set the AWS_SECRET_ACCESS_KEY environment variable: AWS secret access key paired with the access key ID.
  5. Set the AWS_REGION environment variable: AWS region where the Knowledge Base is deployed.
  6. Set the KNOWLEDGE_BASE_ID environment variable: The Amazon Bedrock Knowledge Base ID to query.
  7. Restart your AI client to apply the config.

Related Servers