mcpField
A
Ai Diagram Maker Mcp Server

Ai Diagram Maker Mcp Server

Local
github.com/erajasekarDeveloper Tools10Updated 29w ago

About Ai Diagram Maker Mcp Server

Ai Diagram Maker Mcp Server lets MCP-compatible agents create software diagrams from text, JSON, code, ASCII art, images, and Mermaid. It supports local and hosted use, with diagrams rendered inline and a URL for viewing or editing.

Key features

  1. 01Diagram generation from five input types
  2. 02Five software diagram types
  3. 03Inline rendering with MCP Apps UI
  4. 04Diagram URLs for viewing and editing
  5. 05stdio and HTTP transports

Use cases

  1. 01Turn a written system description into a diagram
  2. 02Visualize JSON, code, or ASCII diagrams
  3. 03Convert a whiteboard image into a clean diagram
  4. 04Render Mermaid definitions as PNG

MCP support

It exposes five diagram-generation tools to MCP clients over stdio or HTTP/Streamable HTTP. Clients can use the hosted endpoint or run the server locally.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"ai-diagram-maker": {
"command": "npx",
"args": [
"-y",
"ai-diagram-maker-mcp"
],
"env": {
"YOUR_API_KEY": "<YOUR_API_KEY>"
}
}
}
}
Run commandnpx -y ai-diagram-maker-mcp
Environment variables to set
YOUR_API_KEYSecretYour API key for the service

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmai-diagram-maker-mcp3.0.6

Tools

5

Saved tool list · updated 3w ago

ParameterTypeRequiredDescription
contentstringYesNatural language description of the diagram to generate. Be descriptive — include components, relationships, data flows, etc. Example: "Create a microservices architecture with API gateway, auth service, user service, and PostgreSQL database"
diagramTypeenumNoPreferred diagram type. Leave blank to let the AI infer the best type from your description.flowchart | sequence | erd | system_architecture | network_architecture | uml | mindmap | workflow
promptstringNoAdditional styling or layout instruction. Example: "Use left-to-right layout with pastel colors"
isIconEnabledbooleanNoSet to true when the user asks to include icons in the diagram.

Details

From the project README

AI Diagram Maker MCP Server

MCP server for AI Diagram Maker — generate beautiful software engineering diagrams directly inside Cursor, Claude Desktop, Claude Code, or any MCP-compatible AI agent.

  • ai-diagram-maker-mcp 🌐 ☁️ - Generate professional software diagrams from plain English descriptions. erajasekar/ai-diagram-maker-mcp MCP server

Features

  • 5 tools covering every input type: natural language text, code, ASCII diagram, images, and Mermaid
  • Inline rendering — diagrams appear directly in the chat using MCP Apps UI, no downloads
  • Diagram URL in responses — open it in your browser to view and edit the diagram
  • 5 diagram types: flowchart, sequence, ERD, system architecture, UML
  • Supports both stdio (local) and HTTP/Streamable HTTP (remote) transports

Contents

Prerequisites

  1. Node.js 18+
  2. An AI Diagram Maker account and API key

Hosted MCP server

The public MCP endpoint is https://mcp.aidiagrammaker.com/mcp (Streamable HTTP). Nothing to install for this option.

Authentication (HTTP)

For remote HTTP clients, send your API key on every request — not via environment variables:

  • X-ADM-API-Key: <your_api_key> (recommended), or
  • Authorization: Bearer <your_api_key>

Use the API key from your AI Diagram Maker account (see Prerequisites).

Remote server JSON example

Merge this into your client’s MCP config (replace the API key placeholder):

{
  "mcpServers": {
    "ai-diagram-maker": {
      "url": "https://mcp.aidiagrammaker.com/mcp",
      "headers": {
        "X-ADM-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Installation

Option A — hosted server (recommended)

Use the remote server JSON example above and wire it into your client using MCP client configuration. No global install.

Option B — run locally with npx

Nothing to install permanently — npx runs the package on demand. The package name is ai-diagram-maker-mcp; append @latest if you want every invocation to resolve the newest release (recommended for one-off runs and claude mcp add).

ADM_API_KEY=your_api_key npx ai-diagram-maker-mcp@latest

MCP client configuration

Cursor
Remote (recommended)

Add to ~/.cursor/mcp.json or Settings → MCP using the remote server JSON example. No environment variables are required for this setup.

Local (stdio)
{
  "mcpServers": {
    "ai-diagram-maker": {
      "command": "npx",
      "args": ["-y", "ai-diagram-maker-mcp@latest"],
      "env": {
        "ADM_API_KEY": "your_api_key_here"
      }
    }
  }
}

Optional: add "ADM_DEBUG": "1" to env for debug logging — see Environment variables.

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "ai-diagram-maker": {
      "command": "npx",
      "args": ["-y", "ai-diagram-maker-mcp@latest"],
      "env": {
        "ADM_API_KEY": "your_api_key_here"
      }
    }
  }
}
Claude Code
macOS
claude mcp add ai-diagram-maker -t stdio -e ADM_API_KEY=<api_key> -- npx -y ai-diagram-maker-mcp@latest
Windows
claude mcp add ai-diagram-maker \
  --command "npx" \
  --args "-y,ai-diagram-maker-mcp@latest" \
  --env ADM_API_KEY=your_api_key_here
HTTP transport (local or self-hosted)

To run an HTTP server yourself (same header-based auth as Authentication (HTTP)):

npx ai-diagram-maker-mcp@latest --transport http

The server listens on $PORT or 3001. Point clients at /mcp and send the API key with each request using the headers above.

Environment variables

VariableRequiredDefaultDescription
ADM_API_KEYYes (stdio only)—Your AI Diagram Maker API key (stdio transport only; remote HTTP clients use headers — see Authentication (HTTP))
ADM_BASE_URLNohttps://app.aidiagrammaker.comOverride for local/staging API; also used as the base for diagram URLs in tool responses
ADM_DEBUGNo—Set to 1, true, or yes to log request parameters from the AI agent and the payload sent to the AI Diagram Maker API. Logs go to stderr. In Cursor, open Output, choose the MCP or ai-diagram-maker channel to read the server logs.

Tools

generate_diagram_from_text

Generate a diagram from a natural language description.

ParameterTypeRequiredDescription
contentstringYesNatural language description of the diagram
diagramTypeenumNoflowchart, sequence, erd, system_architecture, uml
promptstringNoAdditional styling/layout instruction

Example prompts:

  • "Create a microservices architecture with API gateway, auth service, user service, and PostgreSQL database"
  • "Draw a sequence diagram for user login flow with JWT token generation"
  • "adm show the CI/CD pipeline for a Next.js app deployed to Vercel"

generate_diagram_from_json

Convert a JSON structure into a diagram (great for API responses, database schemas, config files).

ParameterTypeRequiredDescription
contentstringYesJSON string to visualise
promptstringNoHow to interpret the JSON
diagramTypeenumNoPreferred diagram type

generate_diagram_from_ascii

Convert ASCII art into a polished diagram.

ParameterTypeRequiredDescription
contentstringYesRaw ASCII art diagram
promptstringNoRendering instructions
diagramTypeenumNoPreferred diagram type

generate_diagram_from_image

Convert a whiteboard photo, screenshot, or any image into a clean diagram.

ParameterTypeRequiredDescription
contentstringYesPublic image URL or base64 data URI
promptstringNoWhat to extract or how to render
diagramTypeenumNoPreferred output diagram type

generate_diagram_from_mermaid

Convert a Mermaid diagram definition to D2 and return a PNG image.

ParameterTypeRequiredDescription
contentstringYesMermaid diagram source (e.g. flowchart, sequenceDiagram, erDiagram)
promptstringNoOptional layout or styling instruction
diagramTypeenumNoPreferred diagram type for the converted output

Trigger keywords

The AI agent will automatically select the right tool when you use phrases like:

  • adm ...
  • ai diagram maker ...
  • create a diagram of ...
  • show me a flowchart / sequence diagram / ERD / architecture ...
  • visualise / draw / diagram ...

Local developer setup

Use these steps to clone the repo, build locally, and run the MCP server with Node.

1. Clone the repository
git clone https://github.com/erajasekar/ai-diagram-maker-mcp.git
cd ai-diagram-maker-mcp
2. Install dependencies
npm install
3. (Optional) Regenerate API client

If you change the AI Diagram Maker OpenAPI spec or config, regenerate the client:

npm run generate
4. Build
npm run build

This compiles TypeScript and builds the MCP app UI into dist/. The server entrypoint is dist/index.js.

5. Run the local MCP server

stdio (default) — for use with Cursor, Claude Desktop, etc.:

ADM_API_KEY=your_api_key node dist/index.js

Or use the npm script:

ADM_API_KEY=your_api_key npm start

HTTP transport — for remote clients or testing (same headers as Authentication (HTTP)):

ADM_API_KEY=your_api_key node dist/index.js --transport http

Or:

ADM_API_KEY=your_api_key npm run start:http

The HTTP server listens on $PORT (default 3001).

6. Use the local server in Cursor

Point Cursor at your built server via Settings → MCP (or ~/.cursor/mcp.json):

{
  "mcpServers": {
    "ai-diagram-maker": {
      "command": "node",
      "args": ["/absolute/path/to/ai-diagram-maker-mcp/dist/index.js"],
      "env": {
        "ADM_API_KEY": "your_api_key_here"
      }
    }
  }
}

Replace /absolute/path/to/ai-diagram-maker-mcp with the actual path to your cloned repo. After changing the config, restart Cursor or reload the MCP servers.

For debug logging, add "ADM_DEBUG": "1" to env — see Environment variables.

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.