Generate AWS Architecture Diagrams Using Kiro CLI (MCP-powered)

sanjid salam

4 min read

·

Dec 6, 2025

--

Listen

Share

Building clean, consistent system diagrams is often harder than building the actual system. You need the right tools, the right icons, and the right styling — and even then, diagrams can look inconsistent across documentation, architecture reviews, or stakeholder presentations.

But with Kiro CLI, generating AWS architecture diagrams becomes dramatically simpler.
Instead of dragging shapes or manually arranging icons, you describe what you want in plain English — and Kiro creates the architecture diagram for you using Model Context Protocol (MCP).

This guide walks you through how to configure Kiro CLI to generate AWS diagrams, and how to build Web App, Three-Tier, and Data Processing architectures.

What is Model Context Protocol (MCP)?

The Model Context Protocol is an open standard that enables AI tools (like Kiro) to securely connect to external data sources, tools, or custom servers.

As Anthropic defines it:

“The Model Context Protocol enables developers to build secure, two-way connections between their data sources and AI-powered tools. Developers can create MCP servers to expose data, or build MCP clients that connect to these servers.”

In our case:
Kiro CLI (the MCP client) connects to AWS Diagram MCP servers, enabling Kiro to generate architecture diagrams directly from your prompts.

Prerequisites

Before starting, ensure:

• You have an AWS Account
• You have permissions to use required services
• You are using macOS / Linux (Windows WSL works too)

Step 1 — Install & Configure Kiro CLI

curl -fsSL https://get.kiro.run | bash

Press enter or click to view image in full size

Login:

kiro login

To verify installation:

kiro --version

Step 2 — Install Python + uv Runtime

pip install uv
uv python install 3.1

Kiro uses uv to run MCP servers smoothly.

Step 3 — Install Graphviz

This is required for rendering diagrams.

brew install graphviz

(Or apt install graphviz on Linux)

Step 4 — Configure MCP Servers for AWS Diagrams

Edit the Kiro config file:

nano ~/.kiro/settings/mcp.json

Add AWS diagram + AWS documentation MCP servers:

{
  "mcpServers": {
    "awslabs.aws-diagram-mcp-server": {
      "command": "uvx",
      "args": ["awslabs.aws-diagram-mcp-server"],
      "env": {
        "FASTMCP_LOG_LEVEL": "ERROR"
      }
    },
    "awslabs.aws-documentation-mcp-server": {
      "command": "uvx",
      "args": ["awslabs.aws-documentation-mcp-server@latest"],
      "env": {
        "FASTMCP_LOG_LEVEL": "ERROR"
      }
    }
  }
}

Save and exit.

Step 5 — Verify Everything Works

kiro-cli

Press enter or click to view image in full size

You should see both AWS MCP servers running successfully.

Moment of Truth — Generate AWS Architecture Diagrams Using Kiro CLI

We’ll generate three diagrams:

1️⃣ Web Application Architecture
2️⃣ Three-tier Architecture
3️⃣ Data Processing Architecture

Kiro can output SVG, PNG, PDF, or raw Graphviz.

To generate any architecture diagram:

kiro generate diagram "<your prompt here>" --output web-architecture.svg

1. Highly Available Web Application Architecture

Prompt Used in Kiro

kiro generate diagram "
Create a detailed AWS architecture diagram using Graphviz for a simple, highly available web application.
Requirements:
- VPC with two Availability Zones.
- Public subnets for ALB + NAT Gateway.
- Private subnets for EC2 + RDS (Multi-AZ).
- ALB distributes traffic to two EC2 instances.
- EC2 communicates with RDS securely.
- Include IGW, security groups, and traffic flow arrows.
- Clearly separate presentation, application, and database tiers.
" --output web-ha-architecture.svg

📌 Output:
A clean AWS-style architecture diagram generated automatically by Kiro CLI.

Press enter or click to view image in full size

2. Three-Tier Architecture with ECS, ALB & Aurora

Prompt Used in Kiro

Generate diagram "
Create a detailed AWS 3-tier architecture diagram. 
VPC:
- Two AZs with public + private subnets.
- IGW + NAT Gateways.Presentation Tier:
- CloudFront
- ALB in public subnetsApplication Tier:
- ECS Fargate running microservices in private subnets
- ECS tasks pull images from ECRData Tier:
- Aurora PostgreSQL Multi-AZ in private subnetsNetworking:
- CloudFront → ALB → ECS → Aurora
- Show SG relationships and subnet groupings.Output goal:
- AWS best practice layout with clear tier separation.
" --output three-tier-architecture.svg

📌 Output:
A fully structured multi-tier AWS architecture diagram.

Press enter or click to view image in full size

3. Data Processing Pipeline Architecture

Prompt Used in Kiro

Generate diagram " Create a data processing pipeline diagram using AWS services.
Four clusters:
1. Data Ingestion – Kinesis Data Streams
2. Data Processing – Lambda, AWS Glue
3. Data Storage – S3 buckets (Raw, Processed, Curated)
4. Analytics – QuickSightInclude:
- Data flow arrows
- IAM role boundaries (optional)
- Regional groupingEnsure AWS best practice visualization.
" --output data-processing-architecture.svg

📌 Output:
A clean multi-stage data pipeline diagram with clear clustering.

Press enter or click to view image in full size

Conclusion

Using Kiro CLI with MCP servers to generate AWS diagrams is a game-changer.
What used to take hours in tools like Lucidchart, Draw.io, or Cloudcraft now takes one clear prompt.

Kiro isn’t trying to replace human-designed architecture diagrams — but it gives you:

• A perfect starting point
• A consistent AWS icon set
• Automated layout + traffic flow
• Faster documentation
• Cleaner communication with teams & stakeholders

As AI tooling evolves, Kiro will become an indispensable partner in how we design, document, and iterate on cloud architecture.