Trabuco

You can use the MCP server to expose Trabuco’s CLI capabilities as structured, AI-friendly tools for your coding assistants. This server enables AI agents to build, inspect, and orchestrate Trabuco projects through well-defined commands and responses, eliminating guesswork and enabling seamless automation across your development workflow.

How to use

You interact with the MCP server by configuring your AI coding agent to communicate via the built-in MCP interface. Once connected, your agent can invoke a set of defined tools to initialize projects, add modules, run health checks, migrate legacy code, and query project metadata. The server standardizes inputs and outputs so your AI assistants receive structured data, making it easier to build, test, and evolve your Java projects.

How to install

Prerequisites: you need Java installed, a compatible shell, and access to the Trabuco CLI. You will also configure your AI agent to connect to the MCP server over standard input/output (stdio). Follow these steps to enable the MCP server workflow.

# Ensure Java is installed (example for a common JDK)
java -version

# Ensure Maven is available (required for building MCP server artifacts)
mvn -version

# Start the MCP server via the CLI so your AI agent can connect over stdio
trabuco mcp

Configuration and client setup

Configure your AI coding agent to connect to the MCP server using a stdio transport. Below are example configurations for common agents. Each entry uses the standard MCP startup command and arguments.

{
  "mcpServers": {
    "trabuco": {
      "command": "trabuco",
      "args": ["mcp"]
    }
  }
}

Security and access

Keep your API keys and credentials secure. Use environment isolation and restrict MCP access to trusted agents and CI systems. Review permissions for each agent and rotate credentials periodically.

Troubleshooting tips

If the MCP server fails to start, verify that the Trabuco CLI is installed and reachable. Ensure there are no port conflicts or blocking policies in your environment. Check for clear error messages in the terminal output when starting the MCP server and address any reported configuration issues.

Examples of supported actions

The MCP server exposes tools to initialize projects, add modules, perform health checks, and query project information. You can request a new Java project with modules, validate the project health, or retrieve current module lists to guide your AI’s code-writing and integration tasks.

Generated project structure overview

When you use MCP-enabled workflows to scaffold or migrate projects, the resulting structure follows a clean modular layout. This helps AI assistants locate models, data access, business logic, and API endpoints quickly, making automated code generation safer and more maintainable.

Notes

The MCP server is designed to be used with the Trabuco CLI to enable AI-assisted development workflows. Start the MCP server with the CLI, then configure your agent to communicate with it using stdio as shown in the examples.

Additional sections

Security best practices, integration notes, and advanced usage patterns are covered in the dedicated MCP server guide available at the repository path mcp_server/README.md. Follow those sections for in-depth guidance on deployment, scaling, and troubleshooting.

Available tools

init_project

Generate a new Java project with specified modules, database, and options

add_module

Add a module to an existing Trabuco project (with dry-run support)

run_doctor

Run health checks on a project and optionally auto-fix issues

get_project_info

Read project metadata from .trabuco.json or inferred from POM

list_modules

List all available modules with descriptions and dependency info

check_docker

Check if Docker is installed and running

get_version

Get the Trabuco CLI version

scan_project

Analyze a legacy Java project's structure and dependencies (no AI required)

migrate_project

Full AI-powered migration of a legacy project (long-running)

auth_status

Check which AI providers have credentials configured

list_providers

List supported AI providers with pricing and model info