Build customizable, well-architected applications on AWS in minutes, not months.
Application Modernization (AppMod) Catalog Blueprints is a comprehensive library of use case-driven infrastructure blueprints and industry-aligned solutions built on AWS Well-Architected best practices. Designed as composable, multi-layered AWS CDK L3 constructs, these blueprints accelerate serverless development and modernization with multiple implementation pathways — from ready-to-deploy solutions to fully customizable building blocks.
Why This Library?
- Use case-driven: Purpose-built for real business problems — AI workflows, document processing, event-driven architectures, web applications
- Multi-layered approach: Infrastructure Foundation → General Use Cases → Industry Solutions — start with proven patterns, customize as needed
- Composable architecture: Mix and match independent components with standardized interfaces
- Enterprise-ready security: Built-in compliance, encryption, least-privilege IAM, and CDK Nag validation
- Multi-language: TypeScript, Python, Java, .NET via JSII
| Approach | Best For | Get Started |
|---|---|---|
| Deploy a Solution | Quick evaluation, proof-of-concepts | Pick from ready-to-deploy examples — deploy in minutes with 3 commands |
| Build Custom | Enterprise integration, tailored applications | Use individual constructs — override defaults, extend and inject custom logic |
| AI-Assisted | Faster development with AI coding assistants | Use the MCP Server + Skills with genAI coding agents (Kiro, Claude Code, Codex etc.) |
Refer to Getting Started for more details.
| Solution | What It Does | Constructs Used |
|---|---|---|
| Customer Service Chatbot | Real-time chatbot with streaming and session management | InteractiveAgent, Frontend |
| Retail Banking Chatbot | Banking chatbot with RAG knowledge base and transaction lookup | InteractiveAgent, BedrockKnowledgeBase, Frontend |
| Retail Banking Chatbot (AgentCore) | High-availability banking bot on AgentCore Runtime | InteractiveAgent, BedrockKnowledgeBase |
| RAG Customer Support | Knowledge-powered Q&A for e-commerce support | BatchAgent, BedrockKnowledgeBase |
| Solution | What It Does | Constructs Used |
|---|---|---|
| Insurance Claims Portal | End-to-end claims processing with web UI | AgenticDocumentProcessing, Frontend |
| Fraud Detection | AI-powered document fraud analysis with risk scoring | AgenticDocumentProcessing |
| Document Summarization | Multi-format summarization with semantic search | BedrockDocumentProcessing |
| Agentic Document Processing | Advanced document processing with AI reasoning | AgenticDocumentProcessing |
| Bedrock Document Processing | Document classification and extraction pipeline | BedrockDocumentProcessing |
| Minimal Document Processing | Zero-config document processing starter | BedrockDocumentProcessing |
| Solution | What It Does | Constructs Used |
|---|---|---|
| Synthetic Dataset Generator | AI-powered synthetic data generation through conversation | InteractiveAgent, BatchAgent, Frontend |
Use individual constructs to build custom applications:
npm install @cdklabs/cdk-appmod-catalog-blueprintsimport { InteractiveAgent } from '@cdklabs/cdk-appmod-catalog-blueprints';
import { Asset } from 'aws-cdk-lib/aws-s3-assets';
new InteractiveAgent(this, 'MyChatbot', {
agentName: 'support-bot',
agentDefinition: {
bedrockModel: { useCrossRegionInference: true },
systemPrompt: new Asset(this, 'Prompt', { path: './prompt.txt' }),
},
});| Construct | What It Does |
|---|---|
| InteractiveAgent | Real-time chatbots with SSE streaming, sessions, and auth |
| BatchAgent | Async AI processing for document analysis |
| BaseAgent | Abstract base for custom agent implementations |
| AgenticDocumentProcessing | Document workflows with AI agents and tools |
| BedrockDocumentProcessing | Document classification and extraction |
| BaseDocumentProcessing | Abstract base for custom document processing |
| BedrockKnowledgeBase | RAG retrieval with access control |
| Frontend | Static web hosting with CloudFront CDN |
| Component | What It Does |
|---|---|
| Network | VPC with subnets and endpoints |
| Observability | Logging, tracing, monitoring with Lambda Powertools |
| DataMasking | PII protection Lambda layer |
┌─────────────────────────────────────────────────────────────────────┐
│ Your Application │
├─────────────────────────────────────────────────────────────────────┤
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Interactive │ │ Batch │ │ Document │ │ Frontend │ │
│ │ Agent │ │ Agent │ │ Processing │ │ (React) │ │
│ └──────┬──────┘ └──────┬──────┘ └──────┬──────┘ └──────┬──────┘ │
├─────────┼────────────────┼────────────────┼────────────────┼────────┤
│ │ │ │ │ │
│ ┌──────▼──────┐ ┌──────▼──────┐ ┌──────▼──────┐ ┌──────▼──────┐ │
│ │ Bedrock │ │ Step │ │ Lambda │ │ CloudFront │ │
│ │ + Cognito │ │ Functions │ │ + S3 │ │ + S3 │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ └─────────────┘ │
├─────────────────────────────────────────────────────────────────────┤
│ Foundation (Network, Observability) │
└─────────────────────────────────────────────────────────────────────┘
| Layer | What It Is | Purpose |
|---|---|---|
| Infrastructure Foundation | Abstract base classes (BaseAgent, BaseDocumentProcessing) |
Standardized interfaces and contracts — extend for custom implementations |
| General Use Case Implementation | Concrete classes (InteractiveAgent, BedrockDocumentProcessing) |
Configurable implementations for common patterns — use directly or extend |
| Industry-Aligned Solutions | Deployable examples (examples/) |
Pre-configured for specific domains (insurance, banking, e-commerce) — deploy as-is or use as reference |
Build complex systems by combining independent, reusable components:
- Independent components with clear interfaces and loose coupling
- Mix and match constructs across different contexts and use cases
- Pluggable strategies for networking, observability, and event handling
- Scalable composition — incremental adoption and gradual modernization
# Configure AWS credentials
aws configure
# OR: export AWS_PROFILE=your-profile-name
# Bootstrap CDK (one-time per account/region)
npx cdk bootstrapClone the repo, pick an example from the catalog, and deploy:
git clone https://github.com/cdklabs/cdk-appmod-catalog-blueprints.git
cd cdk-appmod-catalog-blueprints/examples/chatbot/customer-service-chatbot
npm install && npm run deployThat's it — you now have a well-architected AI chatbot with streaming, authentication, and a React frontend. Swap the path for any other solution.
Import constructs into your own CDK project:
npm install @cdklabs/cdk-appmod-catalog-blueprintsimport { InteractiveAgent } from '@cdklabs/cdk-appmod-catalog-blueprints';
import { Asset } from 'aws-cdk-lib/aws-s3-assets';
new InteractiveAgent(this, 'MyChatbot', {
agentName: 'my-bot',
agentDefinition: {
bedrockModel: { useCrossRegionInference: true },
systemPrompt: new Asset(this, 'Prompt', { path: './prompt.txt' }),
},
});Override defaults, extend base classes, and inject custom logic as needed. See Building Blocks for available constructs.
Use AI coding assistants to scaffold and compose constructs with correct props and dependency wiring. The MCP Server exposes tools for genAI coding agents (Kiro, Claude Code, Codex etc.) via the Model Context Protocol.
See the MCP Server README for setup, or install the appmod-blueprints-builder skill for guided workflows.
All constructs include enterprise-grade security by default:
| Feature | What You Get |
|---|---|
| CDK Nag Integration | Automated security compliance checking |
| AWS Well-Architected | Security, reliability, and performance best practices |
| Encryption | KMS at rest, TLS in transit — always on |
| IAM | Least-privilege, resource-scoped permissions |
| Compliance Reports | Generate with npm test -- --testPathPattern="nag.test.ts" |
- Website — Full documentation and guides
- Construct Hub — API reference
- Examples — Complete deployable solutions
- Building Blocks — Individual construct documentation
- MCP Server — AI-assisted development tools
- Agent Skills — AI assistant skills
We welcome contributions! See CONTRIBUTING.md for guidelines.
These solutions are examples to help you build applications, not supported products. Any applications you build should be thoroughly tested, secured, and optimized according to your security standards before production use.
Apache License 2.0 — see LICENSE for details.