SnapMagic 🎴✨

AI-Powered Trading Card Generation for AWS Events
Generate personalized trading cards and videos using Amazon Bedrock Nova Canvas & Nova Reel

🎯 What is SnapMagic?

SnapMagic creates AI-generated trading cards and videos for AWS events using Amazon Bedrock. Users enter text prompts like "AWS Solutions Architect designing cloud infrastructure" and get personalized trading cards with holographic styling, plus animated videos.

Perfect for: AWS re:Invent, customer summits, partner events, conferences

🏗️ Complete Architecture & Data Flow

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   AWS AMPLIFY   │    │  API GATEWAY    │    │  MAIN LAMBDA    │    │      SQS        │
│                 │    │                 │    │                 │    │                 │
│ • Frontend Host │───▶│ • REST API      │───▶│ • Authentication│───▶│ • Card Queue    │
│ • GitHub Deploy│    │ • CORS Config   │    │ • Job Creation  │    │ • Concurrency   │
│ • Auto Build    │    │ • JWT Validation│    │ • Gallery Load  │    │ • Dead Letter Q │
└─────────────────┘    └─────────────────┘    └─────────────────┘    └─────────────────┘
                                                        │                        │
                                                        ▼                        ▼
┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│    DYNAMODB     │    │       S3        │    │ QUEUE PROCESSOR │    │  AMAZON BEDROCK │
│                 │◀───│                 │◀───│     LAMBDA      │───▶│                 │
│ • Job Tracking  │    │ • File Storage  │    │ • AI Generation │    │ • Nova Canvas   │
│ • Metadata      │    │ • Presigned URLs│    │ • File Upload   │    │ • Nova Reel     │
│ • GSI Queries   │    │ • 7-day Access  │    │ • Status Update │    │ • Text-to-Image │
└─────────────────┘    └─────────────────┘    └─────────────────┘    └─────────────────┘

Complete Data Flow:

1. Frontend (Amplify): User enters prompt via web interface
2. API Gateway: Routes request to Main Lambda with CORS/JWT validation
3. Main Lambda: Creates job in DynamoDB, sends message to SQS queue
4. SQS Queue: Manages Bedrock concurrency (max 2 concurrent generations)
5. Queue Processor Lambda: Picks up job, calls Nova Canvas for AI generation
6. Bedrock Nova Canvas: Generates 1280×720 trading card image from text
7. S3 Storage: Queue Processor saves image, updates DynamoDB with S3 key
8. Gallery Loading: Main Lambda queries DynamoDB, generates 7-day presigned URLs
9. Frontend Display: Shows cards with CSS holographic template styling

Complete AWS Services Stack:

• AWS Amplify: Frontend hosting, GitHub integration, auto-build, environment variables
• API Gateway: REST API endpoints, CORS configuration, request routing
• AWS Lambda (2 functions):
  • Main Lambda: API processing, authentication, job management, gallery loading
  • Queue Processor Lambda: AI generation, file storage, status updates
• Amazon SQS: Card generation queue + Dead Letter Queue for failed jobs
• Amazon DynamoDB: Job tracking table with GSI for device-based queries
• Amazon S3: File storage with presigned URL security and CORS configuration
• Amazon Bedrock: Nova Canvas (images) + Nova Reel (videos) AI models
• CloudWatch: Logging and monitoring for all components

🚀 Quick Deploy

Prerequisites (5 minutes):

1. AWS Account with admin access
2. Bedrock Access: Go to AWS Console → Bedrock → Model Access → Request access to:
   • ✅ Amazon Nova Canvas (image generation)
   • ✅ Amazon Nova Reel (video generation)
3. GitHub Token: Generate here with repo scope
4. Fork this repo to your GitHub account

Deploy Steps:

1. Clone & Configure

git clone https://github.com/YOUR-USERNAME/SnapMagic.git
cd SnapMagic
cp secrets.json.example secrets.json

2. Edit secrets.json (REQUIRED):

{
  "github": {
    "repositoryUrl": "https://github.com/YOUR-USERNAME/SnapMagic",
    "token": "ghp_YOUR_GITHUB_TOKEN_HERE",
    "branch": "main"
  },
  "app": {
    "name": "my-snapmagic-app",
    "region": "us-east-1"
  }
}

⚠️ Critical:

• Use YOUR forked repository URL
• Must be us-east-1 region (Bedrock Nova models)
• GitHub token needs repo permissions

3. Deploy Infrastructure

cd infrastructure
npm install
cdk bootstrap --region us-east-1
cdk deploy SnapMagic-dev --region us-east-1

4. Access Your App

• Frontend URL: Shown in CDK output as AmplifyAppUrl
• Login: demo / demo (or your configured credentials)
• Ready: Generate cards immediately!

🎨 How to Use

Generate Cards:

1. Login with demo credentials
2. Enter prompt: "AWS Lambda expert building serverless applications"
3. Wait 30 seconds for AI generation
4. Download PNG or generate video

Gallery System:

• All your cards saved automatically
• 7-day presigned URL access
• Mobile-friendly display
• Social sharing ready

Video Generation:

• Click "Generate Video" on any card
• Nova Reel animates your trading card
• Download MP4 for social media

🔧 Technical Architecture Details

Two-Lambda Design:

• Main Lambda (SnapMagicAIFunction):

  • Handles all API requests (login, job creation, gallery loading)
  • JWT authentication and validation
  • DynamoDB job creation and queries
  • Presigned URL generation for gallery
  • SQS message sending for card generation
  • 700 concurrent executions for high API traffic

• Queue Processor Lambda (QueueProcessorFunction):

  • Dedicated to AI generation only
  • Triggered by SQS messages
  • Calls Amazon Bedrock Nova Canvas
  • Uploads generated images to S3
  • Updates DynamoDB job status
  • 200 concurrent executions (respects Bedrock limits)

SQS Queue System:

• Main Queue: snapmagic-card-generation-dev
  • 90-second visibility timeout (AI generation time)
  • 20-second long polling for instant pickup
  • Handles Bedrock concurrency limits (max 2 simultaneous)
• Dead Letter Queue: snapmagic-card-generation-dlq-dev
  • 3 retry attempts before moving to DLQ
  • 14-day retention for failed jobs

DynamoDB Design:

• Table: snapmagic-jobs-dev-{timestamp}
  • Primary Key: jobId (UUID)
  • GSI: device-override-index for fast gallery queries
  • Tracks: job status, prompts, S3 keys, timestamps, user metadata
  • Pay-per-request billing for cost efficiency

S3 Configuration:

• Bucket: snapmagic-videos-dev-{account}-{timestamp}
  • Security: Direct URLs blocked, presigned URLs allowed
  • CORS: Configured for GET, HEAD, PUT methods
  • Presigned URLs: 7-day expiration (maximum AWS allows)
  • Auto-cleanup: Files deleted when CDK stack destroyed

Amplify Integration:

• GitHub Connection: Auto-deploys on code changes
• Environment Variables: API Gateway URL injected automatically
• Build Process: Replaces placeholders with actual API endpoints
• Custom Domain: Optional custom domain support

💰 Cost Estimate

For 100 users generating 5 cards each:

• Bedrock Nova Canvas: ~$67 (500 images)
• Lambda: ~$1 (processing)
• DynamoDB: ~$1 (metadata)
• S3: ~$2 (storage)
• Total: ~$71 ($0.71 per user)

Compare to traditional event engagement: $50-100 per person

🛠️ Customization

Event Branding:

• Add logos to frontend/public/logos/ as 1.png, 2.png, etc.
• Update event name in secrets.json
• Logos appear automatically in card template

Authentication:

• Demo mode: demo/demo
• Custom credentials in secrets.json
• JWT-based with configurable expiration

Limits (configurable in secrets.json):

• Cards per user: 5 (default)
• Videos per user: 3 (default)
• Override system for staff unlimited access

🔍 Troubleshooting

Common Issues:

• "Access Denied" for Bedrock: Request Nova Canvas/Reel access in us-east-1
• Cards not generating: Check CloudWatch logs for Lambda errors
• Gallery images not loading: Verify S3 presigned URL configuration
• Frontend shows API errors: Confirm API Gateway URL in Amplify environment

Logs & Monitoring:

• CloudWatch: /aws/lambda/SnapMagic-* log groups
• DynamoDB: Check job status in snapmagic-jobs-* table
• S3: Verify files in snapmagic-videos-* bucket

📊 Architecture Benefits

Why This Design:

• Serverless: Zero server management, infinite scale
• Cost-Effective: Pay only for usage, not idle time
• Reliable: AWS managed services, 99.9%+ uptime
• Secure: IAM roles, presigned URLs, encrypted storage
• Fast: DynamoDB queries, S3 CDN-ready storage

Production Ready:

• Well-Architected Framework compliance
• Infrastructure as Code (CDK)
• Monitoring and logging built-in
• Account-agnostic deployment

🎉 Success Metrics

After deployment, you'll have:

• ✅ AI Card Generation: 30-second creation time
• ✅ Gallery System: Instant loading with presigned URLs
• ✅ Video Creation: 60-second animated videos
• ✅ Mobile Experience: Works on all devices
• ✅ Social Sharing: 7-day accessible links

Transform your AWS events with AI-powered engagement! 🚀

---

Built with ❤️ for the AWS Community
Questions? Check CloudWatch logs or create an issue.