draw io gomcp

    1

    Install and use an MCP server to generate, edit, and export Draw.io diagrams locally using natural language.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    draw io gomcp

    Example session with this skill installed

    Create a sequence diagram for a user login flow including the Frontend, Auth Service, and Database, then save it to ~/diagrams/login.drawio.

    • Read your context and instructions
    • Compiled the draw io gomcp

    I have created the login sequence diagram and saved it to the specified location. I used the draw-io-gomcp tools to initialize the session, define the lifelines for the Frontend, Auth Service, and Database, and map the message flows between them.

    Login Flow Sequence Diagram

    <mxfile host="Electron" modified="2024-03-21T10:00:00.000Z" agent="5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) draw.io/21.1.2 Chrome/106.0.5249.199 Electron/21.4.4 Safari/537.36" version="21.1.2" type="device">
      <diagram id="L-1" name="Page-1">
        <mxGraphModel dx="1422" dy="794" grid="1" gridSize="10" guides="1" tooltips="1" connect="1" arrows="1" fold="1" page="1" pageScale="1" pageWidth="827" pageHeight="1169" math="0" shadow="0">
          <root>
            <mxCell id="0" />
            <mxCell id="1" parent="0" />
            <mxCell id="2" value="Frontend" style="shape=umlLifeline;perimeter=lifelinePerimeter;whiteSpace=wrap;html=1;container=1;collapsible=0;recursiveResize=0;outlineConnect=0;" vertex="1" parent="1">
              <mxGeometry x="100" y="40" width="100" height="300" as="geometry" />
            </mxCell>
            <mxCell id="3" value="Auth Service" style="shape=umlLifeline;perimeter=lifelinePerimeter;whiteSpace=wrap;html=1;container=1;collapsible=0;recursiveResize=0;outlineConnect=0;" vertex="1" parent="1">
              <mxGeometry x="300" y="40" width="100" height="300" as="geometry" />
            </mxCell>
            <mxCell id="4" value="Database" style="shape=umlLifeline;perimeter=lifelinePerimeter;whiteSpace=wrap;html=1;container=1;collapsible=0;recursiveResize=0;outlineConnect=0;" vertex="1" parent="1">
              <mxGeometry x="500" y="40" width="100" height="300" as="geometry" />
            </mxCell>
            <mxCell id="5" value="login(user, pass)" style="verticalAlign=bottom;endArrow=block;entryX=0;entryY=0;shadow=0;strokeWidth=1;" edge="1" parent="1" source="2" target="3">
              <mxGeometry relative="1" as="geometry">
                <mxPoint x="150" y="120" as="sourcePoint" />
              </mxGeometry>
            </mxCell>
            <mxCell id="6" value="validateUser()" style="verticalAlign=bottom;endArrow=block;entryX=0;entryY=0;shadow=0;strokeWidth=1;" edge="1" parent="1" source="3" target="4">
              <mxGeometry relative="1" as="geometry" />
            </mxCell>
            <mxCell id="7" value="userRecord" style="verticalAlign=bottom;endArrow=open;dashed=1;endSize=8;exitX=0;exitY=0.95;shadow=0;strokeWidth=1;" edge="1" parent="1" source="4" target="3">
              <mxGeometry relative="1" as="geometry" />
            </mxCell>
            <mxCell id="8" value="success / token" style="verticalAlign=bottom;endArrow=open;dashed=1;endSize=8;exitX=0;exitY=0.95;shadow=0;strokeWidth=1;" edge="1" parent="1" source="3" target="2">
              <mxGeometry relative="1" as="geometry" />
            </mxCell>
          </root>
        </mxGraphModel>
      </diagram>
    </mxfile>
    

    Next steps

    • Run screenshot_diagram to see a visual preview of the layout.
    • Use edit_diagram to add an alternative flow for "Invalid Credentials".
    • Open the file in the Draw.io desktop app or at diagrams.net to make manual style adjustments.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Stop struggling with manual XML or rigid Mermaid syntax for complex visuals. This skill installs a local MCP server that gives your agent full control over Draw.io (.drawio) files, allowing it to generate, edit, and preview professional diagrams via natural language.

    What it does

    • Automatic installation of the draw-io-gomcp binary for your specific OS and architecture.
    • MCP registration across supported clients like Claude Desktop, Cursor, and Workbuddy.
    • Diagram generation for flowcharts, sequence, class, ER, architecture, and UML diagrams.
    • In-place editing of existing .drawio files, including adding nodes, paths, and pages.
    • Visual verification through screenshot and export tools to preview diagrams before saving.
    • Shape library access to browse and use vendor-specific icon sets and standard libraries.

    How it works

    1. Run the scripts/install.sh command to download the static binary and update your MCP config.
    2. Restart your AI client to enable the 14 new Draw.io specialized tools.
    3. Use natural language to describe a diagram or provide an existing file path for modifications.
    4. The agent uses get_guide to fetch relevant drawing rules and executes the edit or creation.

    Frameworks & tools

    The server is built in Go as a zero-dependency static binary. It interfaces with Draw.io (diagrams.net) file formats and operates over the Model Context Protocol (MCP) using stdio.

    Why this beats prompting it yourself

    Generic prompts often produce broken Mermaid code or invalid XML that requires manual fixing. This skill provides a specialized toolset and a dynamic guide system that ensures the agent follows specific schema rules for layouts, edges, and styles.

    Use cases

    • Generate system architecture diagrams from technical requirement documents.
    • Update legacy documentation by loading and modifying existing .drawio files.
    • Create multi-page UML diagrams for complex software design reviews.
    • Export high-fidelity diagrams for PRs or design specs without leaving the editor.

    Known limitations

    Does not support raster image generation or ASCII-only outputs. Prebuilt binaries require macOS, Linux, or Windows; other systems require a local Go environment for source builds.

    Requires manual client restart after installation. Prebuilt binaries support common OS/arch combinations; others require Go for source builds. No support for raster-only image generation.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 4 days ago

    • Free with an Agensi account

    Requires

    GitHub
    Go
    Git

    Permissions

    Terminal / Shell
    Network Access
    Read Files
    Write Files

    Allowed Hosts

    api.github.com
    github.com
    go.dev
    Listed4 days ago

    What's inside

    Frequently Asked Questions

    Unlike generic AI tools that struggle with XML or simple Mermaid syntax, this skill gives your AI agent specialized tools to manipulate native .drawio files, allowing for professional layouts, multi-page documents, and granular node editing that doesn't break.

    This skill functions through the Model Context Protocol (MCP), meaning it is compatible with any MCP-enabled client including Claude Desktop, Cursor, and Workbuddy on macOS, Linux, and Windows.

    Setup is automated via a provided installation script that detects your system architecture, downloads the zero-dependency Go binary, and updates your MCP configuration files for you.

    Yes, the agent can use the 'Visual Verification' tool to capture screenshots and exports of the diagrams, allowing you to review the output within your AI chat before finalizing the file.

    Absolutely; the skill allows your agent to read existing .drawio files, browse available shape libraries, and perform in-place edits such as adding new paths, nodes, or pages to legacy documentation.

    You receive the draw-io-gomcp server binary, the automated installation scripts, and the dynamic 'Guide' system which provides the agent with the specific drawing rules and schemas needed for high-quality outputs.

    New in LLM Ops