AI made writing the code easy. Running it in production is still hard.
Prod Forge is an open-source reference that shows how to take a system to production: AI-assisted development, quality gates, CI/CD, infrastructure, observability, migrations, and rollback.
The implementation is a simple Todo app on purpose, but the architecture follows patterns used in real production systems, and every major technical decision is documented and explained.
- Safe releases: migrations run as a one-off ECS task before the deploy, a rollback switches to one of the last
three task revisions, and in-flight requests drain on
SIGTERM. Release Management - One trace ID end to end: created in the browser, carried to the backend logs, Sentry and Tempo. Client Observability
- Web and mobile from one core: a React client and an Expo app share one Redux store, API client and set of design tokens. Mobile Client
- A design system with gates: every Storybook story has a visual snapshot and an axe-core accessibility check. Design System
- Infrastructure you can review: dev and prod from one Terraform module, GitHub OIDC instead of static keys, offline
checkovandtflinton every pull request. Environments - Numbers, not promises: RPO about 5 minutes, a cost estimate of about 150 USD a month for dev and 580 USD for prod. Backup and DR, Cost
- AI-assisted development with guardrails:
AGENTS.md, adocs/reference and approved plans in every repository, behind the same quality gates as hand-written code. AI-Assisted Development
Read the guides in this order:
| Step | Guide | Chapters | Time | Question it answers |
|---|---|---|---|---|
| 1 | Backend guide | 16 | ~3 h | How do I build and operate a production-ready API? |
| 2 | Frontend guide | 17 | ~2 h | How do I build, ship, and keep web and mobile clients healthy? |
| 3 | Infrastructure guide | 15 | ~1.5 h | How do I run all of it on AWS with code, not clicks? |
| Layer | Tools |
|---|---|
| Web Client | React · Vite · Redux Toolkit · Tailwind CSS · NX |
| Mobile Client | Expo · React Native · Redux Toolkit |
| Deploy | AWS S3 · CloudFront · ECR · ECS |
| Backend | NestJS · Prisma · PostgreSQL · Redis · Docker |
| Infrastructure | Terraform · AWS · RDS · ElastiCache |
| Observability | Prometheus · Grafana · Loki · Promtail · Tempo |
| Quality | ESLint · Prettier · Husky · Commitlint · CI/CD |
- A reference with explained decisions: chapters explain what was chosen and why.
- A Todo app on purpose: the domain stays trivial so the production practices stay visible.
- Not a boilerplate or a starter kit to fork and ship as is.
- Not a framework: take the patterns that fit your system and leave the rest.
- Identity provider and password management: the API verifies signed JWTs and ships a login-by-email stub, see Authentication and Authorization
- Cursor pagination: list endpoints use offset pagination, see Performance
- Live demo or public deployment
- Feature flags
- A/B testing
Open roadmap: Kubernetes, blue/green deployment and other open items are tracked in the roadmaps of backend, frontend and terraform.
| Repository | Latest version | CI | Docs chapters |
|---|---|---|---|
| backend | 1.0.1 | Runs on pull requests and pushes to main |
16 |
| frontend | 1.1.0 | Disabled on purpose, see the roadmap | 17 |
| terraform | 1.0.1 | Runs on pull requests and pushes to main (offline checks: fmt, validate, tflint, checkov) |
15 |
Contributions of any kind are welcome. Read the guidelines of the repository you want to change: backend, frontend, terraform.
Questions and ideas: Discussions.

