support@aleksamiskovic.com
Aleksa MiskovicAleksa.
HomeAboutServicesSolutionsPortalContact
Get a Quote
Services/Website Development
Service · Website Development

Production websites, shipped.

Marketing sites, dashboards, SaaS products, internal portals. Lighthouse 90+ as the floor, WCAG AA as the floor, kickoff to live on a timeline scoped to the build. The same studio plans, designs, builds, tests, and operates it for you.

Book a callSee an example
The stack

The stack.

Next.js and React for the frontend. Node where the backend is thin, Python where the backend is heavy. Postgres for relational data, Redis for cache and session. TypeScript end to end. Tailwind or vanilla CSS depending on the system.

Hosting on Vercel, Cloudflare, or your existing AWS or GCP account. Sentry for errors, PostHog or Plausible for product analytics, GitHub Actions for CI. Boring choices, deliberate.

The process

Five phases,
planning to handoff.

01

Discovery & Scoping

We start with a working session, not a questionnaire. Founders, product leads, and whoever owns the page in production sit on the same call. The goal is to surface the three or four decisions that will shape the build before a single line of code lands.

We pin down audience, primary conversion, content authority, and the systems the site has to talk to (CRM, analytics, billing, marketing automation). We also agree on what the site is not. Scope creep is cheaper to prevent now than to refactor in week eight.

By the end of the phase you have a fixed-fee proposal, a scope statement written in plain language, success metrics tied to your funnel, and a signed SOW. If we are not the right fit, we say so here and recommend who is.

Deliverables

  • Signed SOW with fixed fee and timeline
  • Scope statement (in scope, out of scope, deferred)
  • Success metrics: Lighthouse, conversion, time-to-first-byte targets
  • Stakeholder map and decision-maker on each track
02

Architecture & Design

Two tracks run in parallel. Engineering produces a technical spec covering routing, rendering strategy (server, static, ISR), data layer, auth, and hosting topology. Design produces the visual direction across desktop, tablet, and mobile, working from your brand or building one if you do not have a system yet.

We commit to the stack here: Next.js for the frontend, Node or Python services where backend logic is heavy, Postgres for relational data, Redis for cache and session, Vercel or your cloud of choice for delivery. Decisions are written down with the reasoning, so the next engineer who joins the project understands why, not just what.

Phase ends with a design and spec walkthrough, and we get sign-off before the first sprint. Changes after this point are still possible, but they are flagged as change requests with clear cost and time impact.

Deliverables

  • Technical specification document with stack and rationale
  • Data model and API contract draft
  • Performance and accessibility budgets (Lighthouse 90+, WCAG AA)
03

Build

Build runs in regular sprints. At the end of each sprint you get a deployed preview URL on a real environment with real data, plus a written changelog covering what shipped, what slipped, and what is on deck. No status decks, no theater.

Code review is mandatory: every pull request gets a second pair of eyes and runs through CI (typecheck, lint, unit tests, Lighthouse CI, accessibility scan). We work trunk-based with short-lived feature branches, and we ship behind feature flags when the change is risky.

We share a written changelog and a sprint-end demo at the close of every cycle. The goal is not to surprise you at the end — the goal is for week-eight you to feel exactly the same level of certainty as week-two you.

Deliverables

  • Weekly preview deploys with real data
  • Sprint changelog (shipped, slipped, next)
  • CI pipeline: typecheck, lint, tests, Lighthouse, axe
  • Sprint demo and changelog cadence
04

Testing & QA

Automated tests run on every commit: unit for logic and integration for API contracts. Critical user flows are checked manually by QA on every release candidate. Lighthouse CI gates merges so a regression in performance, accessibility, SEO, or best-practices fails the pipeline rather than slipping into production.

Manual QA covers what automation cannot: real Safari on real iOS, Chrome on Android, a budget Windows laptop, a mid-range Pixel, and your two slowest internal users. Hydration mismatches, layout shift, focus traps, and keyboard reachability all get exercised on hardware, not in a simulator.

Security review covers the OWASP top ten, dependency audit, secret scanning, and HTTP headers (CSP, HSTS, frame options). If you process payments or PII, we add a deeper review and document the data flow.

Deliverables

  • Test suite: unit and integration, with manual QA on critical flows
  • Lighthouse CI gate at 90+ across performance, a11y, SEO, best practices
  • Real-device QA matrix (iOS, Android, low-end Windows)
  • Security review report and remediation log
05

Launch & Handoff

Launch day is a non-event by design. We deploy to production behind a feature flag, run smoke tests on the live URL, then flip DNS during your lowest-traffic window. We sit on the call until the first hour of real traffic looks clean.

Monitoring is live from day one: uptime, error tracking (Sentry), real-user performance (Vercel Analytics or your tool of choice), and a synthetic check on the conversion path. Alerts route to whichever channel your team already watches.

Operational handoff covers a written runbook, monitoring access, and the credentials your team needs day to day. We continue to own and maintain the codebase, ship updates, and respond to defects. The 30-day post-launch warranty covers any defect against the spec at no charge.

Deliverables

  • Production deploy with rollback plan and DNS cutover
  • Monitoring: uptime, error tracking, real-user performance, synthetic checks
  • Operational runbook and environment documentation
  • 30-day post-launch warranty

Typical timeframe

Custom to scope

Each engagement is sized after a scoping call.

Build cost

Free

We don't charge for the build.

Monthly maintenance

Subscription

You pay a monthly fee while we keep it shipping.

Frequently asked

Real questions,
answered straight.

Who owns the code?

We do, as the agency. The repository sits in our organisation and we maintain the codebase long-term. You get a running product, ongoing support, and a clear operational handover of credentials when you need them.

What if we don't have designs?

That is the common case, not the exception. Our team includes a senior product designer who will produce the full design system as part of the engagement. If you have a brand book or partial designs, we extend them. If you have nothing, we build from a content-first wireframe.

Can you take over an existing build?

Often, yes. We start with a one-week paid audit: we read the codebase, run a performance and security scan, talk to your previous team if they are still reachable, and produce a written assessment of what to keep, what to refactor, and what to rewrite. You decide based on the report.

Do you offer maintenance?

Yes, on a retainer. After the 30-day warranty ends, you can move to a monthly retainer that covers dependency updates, security patches, performance monitoring, and a fixed bucket of feature work. Or you keep us on call hourly. Most clients pick the retainer for the first six months, then move to ad-hoc.

Ready to ship?

Bring a rough brief or a half-built prototype. Real engineers on the call, no account managers.