Get Started

Get started

Overview

The MCP (Model Context Protocol) allows connecting external tools and data to an LLM. SqlBak provides a hosted SqlBak MCP server for interacting with your account.

With SqlBak MCP you can configure and manage backup jobs via an AI chat: create jobs, adjust parameters, troubleshoot errors from logs, and get suggestions based on your configuration.

Connection

In your MCP client, add a remote server with the URL https://mcp.sqlbak.com/mcp.

If your client uses a JSON configuration, add the following server entry:

{
    "mcpServers": {
        "sqlbak-3-mcp": {
            "url": "https://mcp.sqlbak.com/mcp"
        }
    }
}

Connect to the server and follow your MCP client's authorization prompts to sign in to your SqlBak account and authorize access. No API key or local installation is required.

Usage examples

Creating backup / maintenance / restore jobs

Describe in the chat what job to create and what settings you need. Any parameters you don’t specify will use defaults.

Example request:

Create a backup job for nightly backup of all databases and upload them to Amazon S3.

Help with errors

You can ask the model why a job failed. SqlBak MCP will find the related run and logs; the explanation will be based on SqlBak documentation rather than generic advice.

Why did the backup in SqlBak fail yesterday?

Installing the agent and adding a database connection

If you have other MCP tools that run commands on servers, you can ask to install SqlBak and add a database connection. SqlBak MCP will return installation instructions.

Install SqlBak on server my-private-server and connect to the database there. Get the database login and password from the master file.

For more examples, see the Samples page.