Astah Pro MCP

Connects Astah Professional to AI agents through a local MCP server, letting AI retrieve and edit your UML models in real time.

Details

Compatible with

Astah Professional v12 and later (To use this plugin with v11 and earlier, please check the Compatibility.)

Latest Version

0.3.0 (September 13, 2026)

Developer

Takaaki Teshima

Support

Report issues

Access on GitHub Download

Plugin Overview:

Connects Astah Professional to AI agents through a local MCP (Model Context Protocol) server, letting AI retrieve and edit your UML models in real time.

Example Use Cases

• Design a system with AI and have it generate the UML models and diagrams in Astah.
• Ask AI to explain or analyze the models and diagrams in your project.
• Generate source code from UML models – and vice versa.
• Turn hand-drawn sketches into UML diagrams, or AI responses into mind maps.

For demo videos and technical usage instructions, see the full README on GitHub.

Tool Calling ModesNew

Since version 0.3.0, the plugin experimentally implements a programmatic tool calling mode, inspired by the idea described in this Anthropic article. It exposes a far smaller set of tools than the direct tool calling mode, which also remains available. Unless you have a specific reason to choose otherwise, use port 8888.

Supported AI Assistants / Agents

The following agents have been tested by the developer. From the developer’s experience, Claude Code works particularly well for UML modeling, followed by Codex CLI.

• Claude Code
• Codex CLI
• Grok Build
• Antigravity CLI
• Cursor IDE
• Antigravity IDE
• Kiro IDE

For the detailed setup instructions for each AI agent, please follow these instructions on the developer’s documentation – AI agent settings.

Supported Diagram Types

For updates and setup instructions, refer to the developer’s README on GitHub.

Installation Instructions:

1. Download the latest .jar from the release page.
2. In Astah, navigate to [Plugin] -> [Installed Plugins].
3. In the dialogue box, click [Install], then select the downloaded .jar file.
4. Restart Astah. If the installation is successful, you will see a new [mcp] tab below the Diagram Editor.
   
5. The MCP server uses Streamable HTTP transport. When configuring your AI assistant/agent, use the following:
   • Use 127.0.0.1 — do not use localhost (connections to localhost will fail).
   • Programmatic tool calling mode: port 8888 → http://127.0.0.1:8888/mcp
   • Direct tool calling mode: port 18888 → http://127.0.0.1:18888/mcp
6. Each port can be changed with an environment variable if it conflicts with something else on your machine. Remember to update your AI agent settings to the same port number.
   • Programmatic tool calling: ASTAH_PRO_MCP_PORT_FOR_PROGRAMMATIC
   • Direct tool calling: ASTAH_PRO_MCP_PORT_FOR_DIRECT

For full configuration instructions for each AI assistant/agent, refer to the  the plugin developer’s documentation below:

• AI agent settings
• Changing the Port Numbers

How to use

1. Launch Astah Professional.
   (Note: Start Astah before your AI assistant, as some AI assistants/agents attempt to connect to the MCP server immediately upon startup.)
2. Start the AI assistant/agent.
3. Each time an AI assistant/agent establishes a new session with the MCP server, a connection dialog will appear in Astah. Review the details and click “Connect”.
   
4. You can now send prompts to the AI assistant/agent and begin AI-powered modeling directly.

To disable the plugin, click [Plugin] > [Installed Plugins], select the Astah Pro MCP entry in the plugin list dialog, click [Disable], and restart Astah.

Example Prompts

The developer’s README shows the results of the following prompts, run in sequence against a single Astah project:

1. Design a simple library management system and draw a class diagram in Astah to illustrate its structure.
2. Based on the contents of the Astah model, draw a sequence diagram to illustrate the behavior of borrowing a book.
3. Based on the contents of the Astah model, draw an activity diagram to illustrate the behavior of returning a book.
4. Based on the contents of the Astah model, draw a state machine diagram to illustrate the state transitions of a book.

   
   

See the resulting diagrams from Claude Code, Codex CLI and Grok Build in the demo section of the README.

Important Notes:

• This project is currently experimental. The design and implementation may undergo breaking changes.
• This MCP server only connects to AI agents running on the same machine as Astah Professional.
• Astah project data and logs will be shared with the connected AI assistant/agent. If your project contains sensitive or confidential information, consider opting out or applying appropriate safeguards.
• Since the MCP server can modify your models and diagrams, we strongly recommend using version control (e.g., Git) or saving backups frequently.
• Some model/diagram properties may not be viewable or editable through the current API. Use Astah’s GUI to modify such elements.
• Token-saving techniques are not implemented. We recommend using a flat-rate AI service plan to avoid unexpected usage charges.
• As with code, AI-generated output may include errors or misinterpretations of your UML models.
• You may operate Astah through AI services or external applications only if you hold a valid license and use it for your own purposes. Allowing a non-licensed third party to operate Astah via such an agent is prohibited. Please refer to the official FAQ about the AI service usage.

Plugin Compatibility Table

Plugin Version	Update	Compatible with	Download
0.2.6	2026.7.06	Astah Professional Version 11.0
0.1.5	2025.11.18	Astah Professional Version 10.1

About Support:

This plugin is developed and maintained by a user of Astah. For feature requests or bug reports, please contact the developer directly via the plugin’s GitHub Issues page. Because the project is in an experimental phase and may introduce breaking changes, the developer is not currently accepting pull requests.

For technical support, please directly contact the plug-in developer.