5.14. Use the MCP Server for your AI Workflows

This guide explains how to connect to the Speedgain for Databases MCP server using the Streamable HTTP protocol.

5.14.1. Connection Overview

The MCP server is exposed through the NginX Reverse Proxy. This allows you to access the server using standard HTTP/HTTPS ports without needing to expose internal Docker ports.

Parameter Value

Transport Type

Streamable HTTP

Connection URL

http://<your-server-address>/Speedgain_for_Databases_Mcp/

Authentication

Not required

5.14.2. Connection verification using the MCP Inspector

The MCP Inspector is a web-based tool used to verify that the server is running correctly and to explore available tools.

  1. Open the MCP Inspector in your browser.

  2. Set Transport Type to Streamable HTTP.

  3. In the URL field, enter your connection URL:

  4. Set Connection Type to Via Proxy.

  5. Leave the Authentication and Proxy Session Token fields as is.

  6. Click Connect.

Once connected, you should see the green status indicator and the version tag.

5.14.3. Client Integration (e.g., Claude Desktop)

To use the Speedgain MCP server directly within Claude Desktop or any other AI Tool, you need to configure it as a remote resource.

Add the following entry to your claude_desktop_config.json file. This tells Claude to connect to your running Docker service via the HTTP endpoint.

{
  "mcpServers": {
    "speedgain": {
      "url": "http://localhost/Speedgain_for_Databases_Mcp/"
    }
  }
}
Note
Ensure that your Docker container and NginX are running before starting your MCP client.

5.14.4. Troubleshooting

  • Connection Error: Ensure "Via Proxy" is selected in the Inspector settings.

  • 404 Not Found: Double-check that the URL ends with a trailing slash (/).

  • Logs: If the connection fails, check the s4dbs_mcp Docker container logs for incoming request errors.