Azure Diagram MCP Server
An MCP server that generates professional infrastructure diagrams using the Python diagrams DSL, with first-class Azure support and GitHub Copilot integration for natural language diagram generation.
README
Azure Diagram MCP Server
An MCP server for generating professional infrastructure diagrams using the Python diagrams DSL — with first-class Azure support and GitHub Copilot SDK integration for natural language diagram generation.
graph LR
A[AI Assistant] -->|Natural Language| B[MCP Server]
B -->|Python DSL| C[Diagrams + Graphviz]
C -->|PNG| D[MCP Apps Viewer]
B -->|Security Scan| E[AST + Bandit]
E -->|Pass| C
Getting Started
Step 1 — Install Prerequisites
| Dependency | Install | Verify |
|---|---|---|
| uv | astral.sh/uv | uv --version |
| Python 3.12+ | uv python install 3.12 |
python3 --version |
| Graphviz | brew install graphviz / apt install graphviz / graphviz.org |
dot -V |
⚠️ Graphviz is required. Without it the MCP server will fail to start. Verify with
dot -Vbefore proceeding.
Step 2 — Verify the Server Starts
Run the server directly to confirm everything works:
uvx microsoft.azure-diagram-mcp-server
You should see a message confirming the server is installed and ready. The server is an MCP stdio server — it's designed to be launched by an MCP client, not run directly. If it fails to install, check that Graphviz is installed (dot -V).
Step 3 — Connect to Your AI Host
Pick one of the methods below to register the server with your AI host.
Copilot CLI
-
Start a Copilot CLI session:
copilot -
Inside the session, run the slash command:
/mcp add -
Fill in the form (use Tab to move between fields):
Field Value Name azure-diagramType LocalCommand uvx microsoft.azure-diagram-mcp-server -
Press Ctrl+S to save.
-
Verify with
/mcp show azure-diagram— status should show ✓ Connected.
The config is saved to
~/.copilot/mcp-config.json. You can also edit that file directly:{ "servers": { "azure-diagram": { "type": "local", "command": "uvx microsoft.azure-diagram-mcp-server", "tools": ["*"] } } }
VS Code (one-click)
Or add manually to your VS Code settings.json:
{
"mcp": {
"servers": {
"azure-diagram": {
"command": "uvx",
"args": ["microsoft.azure-diagram-mcp-server"],
"env": {
"FASTMCP_LOG_LEVEL": "ERROR"
}
}
}
}
}
Docker
docker build -t microsoft/azure-diagram-mcp-server .
{
"mcp": {
"servers": {
"azure-diagram": {
"command": "docker",
"args": ["run", "--rm", "-i", "--env", "FASTMCP_LOG_LEVEL=ERROR",
"microsoft/azure-diagram-mcp-server:latest"]
}
}
}
}
Features
| Feature | Description |
|---|---|
| ☁️ Azure-First | 100+ Azure service icons — App Service, Functions, Cosmos DB, AKS, and more |
| 🌐 Multi-Cloud | AWS, GCP, Kubernetes, on-premises, and custom icon support |
| 📊 Multiple Types | Architecture, sequence, flow, class, K8s, and custom diagrams |
| 🔒 Security Scanning | AST + Bandit code analysis before every execution |
| 🖼️ MCP Apps Viewer | Interactive diagram viewer with pan, zoom, download, and dark/light theme |
| 🤖 Copilot SDK | Natural language diagram generation via GitHub Copilot SDK |
Architecture
graph TB
subgraph "GitHub Copilot"
CLI[Copilot CLI]
VS[VS Code Copilot]
end
subgraph "MCP Server"
S[server.py<br/>FastMCP]
DT[diagram_tools.py<br/>Generation + Examples]
SC[scanner.py<br/>AST + Bandit]
V[viewer/app.html<br/>MCP Apps Viewer]
end
subgraph "Copilot SDK Layer"
CC[copilot_client.py<br/>DiagramCopilotClient]
AG[Custom Agent<br/>azure-diagram-architect]
end
CLI & VS -->|MCP stdio| S
CC -->|MCP local| S
AG --> CC
S --> DT
S --> V
DT --> SC
SC -->|Pass| DT
DT -->|Python DSL| GV[Graphviz → PNG]
MCP Tools
| Tool | Description |
|---|---|
generate_diagram |
Execute Python diagram code with security scanning and timeout. Pre-imports all providers — just start with with Diagram(...). |
refresh_diagram |
Regenerate a diagram from updated code (app-only, used by the MCP Apps viewer). |
get_diagram_examples |
Get example code by type: azure, sequence, flow, class, k8s, onprem, custom, or all. |
list_icons |
Discover available icons by provider and service. Filter with provider_filter and service_filter. |
Recommended Workflow
sequenceDiagram
participant User
participant Copilot as GitHub Copilot
participant MCP as MCP Server
participant App as MCP Apps Viewer
User->>Copilot: "Create an Azure web app diagram"
Copilot->>MCP: list_icons(provider_filter="azure")
MCP-->>Copilot: Available icons
Copilot->>MCP: get_diagram_examples(diagram_type="azure")
MCP-->>Copilot: Example code
Copilot->>MCP: generate_diagram(code="...")
MCP-->>Copilot: PNG + structuredContent
Copilot->>App: Render diagram in viewer
App-->>User: Interactive diagram with pan/zoom
MCP Apps Viewer
The server includes an interactive MCP Apps viewer that renders diagrams inline in VS Code and the Copilot CLI. When generate_diagram returns a result, the viewer is automatically displayed.
graph LR
subgraph "MCP Server"
GD[generate_diagram] -->|CallToolResult| SC[structuredContent<br/>status + imageData]
SC --> META["_meta.ui.resourceUri<br/>ui://diagram-viewer/app.html"]
end
subgraph "MCP Apps Viewer"
META --> V[Interactive Viewer]
V --> PAN[Pan & Drag]
V --> ZOOM[Zoom In/Out]
V --> DL[Download PNG]
V --> THEME[Dark/Light Theme]
V --> FIT[Fit to View]
end
| Feature | Control |
|---|---|
| Pan | Click and drag |
| Zoom | Mouse wheel, + / - keys |
| Fit to view | 0 key or toolbar button |
| Download | Toolbar download button |
| Theme | Toggle dark/light in toolbar |
The viewer is served as an MCP resource at ui://diagram-viewer/app.html and receives the diagram as base64-encoded PNG via structuredContent.imageData.
Quick Example
from diagrams import Diagram
from diagrams.azure.compute import AppServices, FunctionApps
from diagrams.azure.database import CosmosDb
from diagrams.azure.network import ApplicationGateway
with Diagram("Azure Web Architecture", show=False):
gateway = ApplicationGateway("Gateway")
app = AppServices("App Service")
functions = FunctionApps("Functions")
db = CosmosDb("Cosmos DB")
gateway >> app >> db
gateway >> functions >> db
Copilot SDK Integration
The server includes a GitHub Copilot SDK client that provides a natural language interface to diagram generation — describe what you want and the Copilot-powered architect generates it.
graph LR
U[User Prompt] --> CC[DiagramCopilotClient]
CC -->|Creates Session| CS[CopilotClient]
CS -->|Connects| MCP[Diagram MCP Server]
CS -->|Uses| AG[azure-diagram-architect<br/>Custom Agent]
MCP -->|Returns| IMG[PNG Diagram]
Interactive CLI
uv run microsoft.azure-diagram-copilot
Programmatic Usage
import asyncio
from microsoft.azure_diagram_mcp_server.copilot_client import DiagramCopilotClient
async def main():
async with DiagramCopilotClient(model="gpt-4.1") as client:
client.on_delta(lambda delta: print(delta, end="", flush=True))
client.on_idle(lambda: print())
await client.generate(
"Create a 3-tier Azure architecture with App Gateway, "
"App Service, and Cosmos DB"
)
asyncio.run(main())
BYOK (Bring Your Own Key)
Use your own LLM provider — no Copilot subscription required:
| Variable | Description |
|---|---|
DIAGRAM_COPILOT_PROVIDER_TYPE |
openai, azure, or anthropic |
DIAGRAM_COPILOT_BASE_URL |
API endpoint URL |
DIAGRAM_COPILOT_API_KEY |
API key |
DIAGRAM_COPILOT_WIRE_API |
completions or responses |
DIAGRAM_COPILOT_MODEL |
Model override (default: gpt-4.1) |
DIAGRAM_COPILOT_AZURE_API_VERSION |
Azure API version (default: 2024-10-21) |
export DIAGRAM_COPILOT_PROVIDER_TYPE=azure
export DIAGRAM_COPILOT_BASE_URL=https://your-resource.openai.azure.com
export DIAGRAM_COPILOT_API_KEY=your-api-key
uv run microsoft.azure-diagram-copilot
Resumable Sessions
client = DiagramCopilotClient(session_id="my-project-diagrams")
await client.start()
await client.generate("Create an Azure web app diagram")
# Resume later
await client.resume("my-project-diagrams")
await client.generate("Add a Redis cache to the previous diagram")
await client.stop()
Development
# Setup
uv sync --group dev
# Test (140 tests, 9 skip without Graphviz)
uv run pytest tests/ -v
# Lint + format
uv run ruff check microsoft/ tests/
uv run ruff format --check microsoft/ tests/
# Type check
uv run pyright
# Coverage
uv run pytest --cov=microsoft --cov-report=term-missing tests/
See AGENTS.md for comprehensive contributor documentation covering architecture, conventions, testing patterns, CI/CD, and the GitHub Pages docs site.
Documentation
📖 microsoft.github.io/diagrams-mcp-server — Full documentation built with VitePress, deployed via GitHub Pages.
cd docs-site && npm install && npm run docs:dev # Local dev server
License
This project is licensed under the MIT License — see the LICENSE file for details.
Contributing
This project welcomes contributions and suggestions. See AGENTS.md for the full development guide.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.