Redshift Administrative MCP Server & Web UI

What is Redshift Administrative MCP Server & Web UI?

This project provides a Model Context Protocol (MCP) server and a Flask web UI for Amazon Redshift administration. The MCP server lets AI agents, coding assistants, and automated tools execute administrative SQL scripts and custom queries on Redshift, using credentials fetched from AWS Secrets Manager. The web UI offers a browser-based interface for connecting and running scripts. Both components are aimed at database administrators and developers who manage Redshift clusters.

How to use Redshift Administrative MCP Server & Web UI?

Install the package with uv pip install -e .. For the MCP server, run uvx redshift-admin-mcp-server with required arguments like --host, --database, --secret-arn, and --aws-region. For the Flask Web UI, set environment variables (FLASK_APP, SECRET_KEY) and run flask run. Connect the MCP server to clients such as Claude Desktop, Cline, Cursor, or a custom Python client via stdio transport.

Key features of Redshift Administrative MCP Server & Web UI

• Connects to Redshift using AWS Secrets Manager credentials.
• Exposes MCP tools for admin scripts and custom SQL queries.
• Outputs results in JSON, CSV, or formatted table.
• Provides schema discovery for MCP clients.
• Includes a Flask web UI for browsing and executing scripts.

Use cases of Redshift Administrative MCP Server & Web UI

• AI agents monitoring and troubleshooting Redshift lock waits or performance.
• Developers running admin scripts (e.g., lock_wait) from an IDE or terminal.
• Database administrators scheduling automated health checks via MCP clients.
• Teams using a web dashboard to explore and execute maintenance queries.

FAQ from Redshift Administrative MCP Server & Web UI

How does the MCP server authenticate to Redshift?

It uses AWS Secrets Manager. You must provide a Secret ARN that contains username and password, along with the Redshift host, database, and AWS region. Credentials can be supplied via command-line arguments or environment variables.

What are the runtime requirements?

Python 3.10 or higher and the uv package manager. An AWS account with permissions to connect to Redshift and access Secrets Manager is also required.

Where is the data stored?

All data resides in the Amazon Redshift cluster you connect to. The MCP server and web UI do not store any data locally; they only run queries and display results.

Can I connect interactively without startup parameters?

Yes. If required parameters (host, database, secret_arn) are omitted, the MCP server starts without a connection. You can then use the connect_redshift tool interactively via an MCP client.

What transport does the MCP server use?

It communicates over stdio transport, which is compatible with all major MCP clients (Claude Desktop, Cline, Cursor, etc.).