Uses — the gear behind the missions
Developers publish a /uses page listing what they actually build with. This is mine: no aspirational logos, just the tools running in production for me and my clients — including the exact stack serving the page you are reading.
Four slots, nothing decorative
What the apps are made of
One stack, chosen so one senior developer can go from idea to production fast — and so clients inherit code that any developer can pick up.
- React, Next.js + TypeScript — every app, this site included — typed end to end
- Node.js, Java + Spring Boot — APIs and services, REST or GraphQL
- Payload CMS — the admin panel clients get; lives inside the same app
- PostgreSQL / SQLite — right-sized per project, with MongoDB and Elasticsearch where they fit
- Stripe — payments and subscriptions when a product charges money
- Jest, Cypress + Playwright — tests that run on every push
AI and automation
AI-assisted development is how fixed prices stay fixed, and agents plus automations save clients the most hours.
- Claude + Claude Code — pair programmer on every build — including this page
- OpenAI — where a different model fits the job better
- n8n — self-hosted automations that glue client systems together
- RPA — for the systems that never got an API
How it reaches production
No hand deploys, ever. A push to the main branch is the only road to production, and old releases stay warm for one-command rollbacks.
- GitHub + GitHub Actions — lint, typecheck, build — fail any step, ship nothing
- Docker — the same container from laptop to production
- AWS, GCP or a VPS — EC2, Elastic Beanstalk, S3 or Google Cloud; this site runs on PM2 + nginx on my own VPS
- My own uptime monitor — a probe every minute; process gaps count as downtime
- webaudit — my audit engine, open-sourced — the same checks as /scan
Running the day to day
The tools around the code: where the work is planned, written and talked about.
- VS Code + IntelliJ — the editors everything above gets built in
- Git + GitHub — every line versioned, every change reviewed
- Jira or Linear — whatever board your team already uses
- WhatsApp — where most clients actually answer
The site you’re on is the demo
thejavadeveloper.com is built exactly the way I build for clients: one Next.js app with the Payload admin inside it, deployed only by the pipeline, monitored by its own uptime probe. The terminal on the right is its real loadout, not a brochure.
Want the receipts in motion — the pipeline replayed, the admin collections, the live monitor? That’s the process page.
$ cat /srv/thejavadeveloper/loadoutapp next.js + payload cms, one processdb sqlite file · survives every deploydeploys github actions → ssh · zero by handruntime pm2 + nginx on my own vpsedge nginx + let's encrypt tlsmonitor self-probe, 60s · see the home page▲ rollback: previous release kept warm
Point this loadout at your problem
/uses pages are a developer tradition — hundreds more at uses.tech.