Obtain the Liquibase AI Changelog Generator Docker image

Last updated: June 15, 2026

The Liquibase AI Changelog Generator is distributed as a Docker image, providing a consistent and isolated environment regardless of your operating system. To get started, you'll authenticate with Liquibase's Docker registry, pull the server image to your local machine, and run it as a container. This guide walks you through each step of the installation process.

Before you begin

Ensure Liquibase Secure is installed on your machine.
Download the AI Changelog Generator Docker image which is publicly available on Docker Hub.
Windows: Confirm you have Docker Desktop installed and open so the Docker daemon is running.
Linux: Confirm you have the Docker daemon running by running the
sudo systemctl start docker command.
Liquibase Secure is installed.

Procedure

Pull the latest snapshot of the AI Changelog Generator by running this command in the CLI:

docker pull liquibase/liquibase-mcp-changelog-server:latest

Run the server container with your database configuration.

Use docker run to start the server. By default, the server generates XML output. To generate formatted SQL for your specific database, pass environment variables using the -e flag in Option 2.

Choose your configuration:

Option 1: XML output (default)

Default configuration (XML output):

docker run liquibase/liquibase-mcp-server:latest

Option 2: Formatted SQL output Set LIQUIBASE_MCP_SQL_DATABASE to match your actual target database to get the most accurate SQL output.

Supported values are:postgresql,mysql,mariadb,oracle,mssql,db2,h2,sqlite.

Formatted SQL output targeting PostgreSQL:

docker run \
  -e LIQUIBASE_MCP_OUTPUT_FORMAT=formatted-sql \
  -e LIQUIBASE_MCP_SQL_DATABASE=postgresql \
  liquibase/liquibase-mcp-server:latest

Formatted SQL output targeting MySQL:

docker run \
  -e LIQUIBASE_MCP_OUTPUT_FORMAT=formatted-sql \
  -e LIQUIBASE_MCP_SQL_DATABASE=mysql \
  liquibase/liquibase-mcp-server:latest

Important: The LIQUIBASE_MCP_SQL_DATABASE setting only applies when LIQUIBASE_MCP_OUTPUT_FORMAT is set to formatted-sql. It has no effect on XML output. For details on database-specific SQL differences, see Database Targeting for Formatted SQL.

Further learning: For a full explanation of both output formats and guidance on which to choose, seeOutput Formats in the Liquibase AI Changelog Generator.

For details on database-specific SQL differences, seeDatabase Targeting for Formatted SQL.

Next steps

Now that you've obtained the Liquibase AI Changelog Generator image from Docker, you're ready to Configure the server image in VS Code.