Production-Ready Serverless Synthetic Data Generator
Plot Palette Dataset »
Generate synthetic training data at scale using AWS Bedrock foundation models. Leverage ECS Fargate Spot instances for up to 70% cost savings with automatic checkpoint recovery. Configure jobs, upload seed data, and monitor progress through a modern React web interface.
** THIS REPO IS IN ACTIVE DEVELOPMENT AND WILL CHANGE OFTEN **
├── frontend/ # React web application
├── backend/ # Lambda functions + ECS workers
├── docs/ # Documentation
└── tests/ # Centralized test suites
- Node.js v24 LTS (for frontend and scripts)
- Python 3.13+ (for backend Lambda functions)
- AWS CLI configured with credentials (
aws configure) - AWS SAM CLI for serverless deployment
npm install # Install dependencies
npm run deploy # Deploy backend to AWS
npm start # Start frontend dev server
npm run check # Run all lint and testsnpm run deployThe deploy script prompts for configuration:
| Prompt | Description |
|---|---|
| Stack Name | CloudFormation stack name (default: plot-palette) |
| AWS Region | Deployment region (default: us-east-1) |
| Environment | development or production |
Defaults are saved to .env.deploy and shown in brackets on subsequent runs.
See docs/README.md for full documentation.
Apache 2.0 - See LICENSE for details.