Web Engineering

Web platforms that hold up in production

Enterprise web applications, headless frontends and design systems, engineered for performance, accessibility and the search engines and AI crawlers that read them.

The web is your most public engineering artifact. We build server-rendered, measurable, accessible platforms on React and Next.js, the same stack this site runs on, and treat Core Web Vitals, SEO and accessibility as engineering requirements, not afterthoughts.

Server-renderedCore Web VitalsAccessible

The call we usually get

A relaunch went wrong, or never shipped: the agency delivered a design, the site is slow, rankings dropped, the CMS fights every editor, and nobody can say why the funnel leaks. The business needs the web platform treated like production software, with performance budgets and a deploy pipeline that does not scare anyone.

When a website needs to be engineered, not assembled

Most business sites are assembled from a builder or a bought template, and they work until growth makes them slow, hard to rank, or impossible to change. A site needs real engineering when speed and Core Web Vitals affect what you earn, when accessibility is a legal and reputational requirement, when search and AI answer engines decide whether buyers find you, and when the site is the front door to a product rather than a brochure. When a template genuinely fits a simple marketing presence, we say so and point you to the lighter path instead of overbuilding.

  • Performance and Core Web Vitals move conversion and ranking, not just a score
  • Accessibility is a requirement you have to meet, not a nice-to-have
  • Search and AI answer engines decide whether buyers find you at all
  • The site is the front door to a product, with logins, data and real functionality

How the web stack fits together

Visitor & edge / CDN
Next.js app (SSR / ISR)
API & data layer
CMS / commerce
Monitoring & Web Vitals

An engineered web stack: rendered at the edge, served fast, measured against performance budgets, and wired to your content and commerce systems.

Where teams put this to work

Concrete situations an engineered web build is made for, across different teams and stages.

Mittelstand

The marketing site loads slowly on mobile and leads drop off before the form.

We rebuild it server-rendered with a real performance budget, so it stays fast on a phone and Core Web Vitals stop costing you conversions.

Marketing leads

Competitors outrank you and the new AI answer engines never cite your pages.

We engineer technical SEO and structured data into the build so search and AI crawlers can read every page cleanly, in every locale.

Compliance-bound teams

You need to meet accessibility obligations and the current site quietly fails them.

We build to WCAG with keyboard, contrast and screen-reader testing in the pipeline, so accessibility is verified, not assumed.

Product teams

You need a real web app with logins and dashboards, not another brochure page.

We build it on Next.js with typed contracts to your backend, so the public site and the product behind the login share one clean codebase.

Commerce teams

The storefront theme is bloated and every change risks breaking checkout.

We move you to a headless frontend on your commerce backend, so the catalog stays fast and the team ships changes without fear.

Agencies

You won the build but lack senior frontend capacity to deliver it under your name.

We deliver it as a bounded work package under NDA, built under your brand and invisible to your client, with the code passing through you so you keep the relationship.

Who this is for

Marketing and commerce

Marketing and digital teams

Editors get a CMS that helps instead of fighting back; the brand gets a design system; releases stop being events.

A CMS and design system you own, EU-hosted, GDPR-grade, handed over for your team to run.

Commerce leaders

Storefronts and portals where speed is revenue: measured, budgeted, monitored on the templates that matter.

Performance budgets and a stack you control, no lock-in, measured against revenue.

Request a quote

Engineering leadership

Platform CTOs

Headless architectures that decouple frontend pace from backend stability, with contracts both sides can rely on.

Headless contracts and your code, model-neutral, EU-hosted, no single-vendor dependency.

Scope the pilot

How we engineer the web

Performance, accessibility and search treated as engineering requirements from the first commit.

Performance budgets

Core Web Vitals tracked against a budget in CI, so a slow change is caught before it reaches your visitors.

Accessibility built in

Built to WCAG with keyboard, contrast and screen-reader checks in the pipeline, so accessibility is verified, not assumed.

SEO and GEO by architecture

Server rendering, structured data and clean semantics so search engines and AI answer engines can read every page in every locale.

Headless and component-driven

A headless frontend on Next.js and React with a reusable component system, so the site stays consistent and quick to extend.

Type-safe end to end

Typed contracts from your backend and content source to the UI, so content and feature changes ship without breaking pages.

Yours to own

Your repository, your hosting, your content model and your IP, handed over with documentation and runbooks so any developer can take it on.

What we build

Enterprise web applications

Customer portals, dashboards and self-service surfaces on React and Next.js.

  • Server-side rendering for SEO and first paint
  • Type-safe from API to component
  • Authentication and roles done properly

Headless frontends

Frontends for Pimcore, Vendure and headless CMS backends.

  • The content model respected, not fought
  • Preview workflows for editors
  • API contracts that survive redesigns

Design systems

Component libraries that keep ten teams shipping one brand.

  • Tokens instead of hardcoded values
  • Accessibility baked into the components
  • Documentation designers actually use

Performance and SEO engineering

Audits and fixes with measured before and after: Core Web Vitals, crawlability, structured data.

  • Performance budgets enforced in CI
  • Structured data for search and AI answers
  • Findings ranked by business impact

How web work ships

Everything ships behind preview deployments: every change is visible on a URL before it is live, and performance budgets fail the build before they fail your users.

How we build it

  1. 1Audit and architecture: measure what exists, decide what carries over
  2. 2Build in weekly slices with preview deployments and performance budgets in CI
  3. 3Launch with monitoring, then a measured handover: your team deploys the last week with us

What you get

  • A platform passing Core Web Vitals on the templates that matter
  • Structured data and server rendering that search engines and AI crawlers can read
  • A deploy pipeline and documentation your team runs without us
Loading diagram...

Reference architecture: CDN / Edge, Next.js SSR, API layer, CMS / Commerce, CI: tests + performance budgets, Monitoring + Web Vitals

How we work

From the first call to a system running in production, and supported after.

    01

    Discovery and planning

    We map the process, the constraints and the people who use it, then agree the scope and the shape of the system before any code is written.

    02

    Architecture and design

    We design the domain model, the data and the interfaces, and write the decisions down so the system stays understandable as it grows.

    03

    Build, reviewed and tested

    We build in small, reviewed increments, type-safe and covered by tests that run on every change, so regressions are caught before you see them.

    04

    Infrastructure and release

    We deploy into your cloud and your accounts through an automated pipeline, with releases you can repeat and roll back without drama.

    05

    Observe and monitor

    We ship logging, metrics and alerts from day one, so we see problems early, often before your users report them.

    06

    Support and iterate

    After launch we fix, extend and harden on a cadence that fits you, with full handover so you are never dependent on us to keep running.

How we run the engagement

Agile sprints

When scope will evolve: we ship in short increments and you steer priorities as the product takes shape.

Fixed work package

When scope is defined and you need a firm price: a contract with result responsibility and a fixed deliverable.

Ongoing partnership

When the system is live and growing: a retainer for changes, support and new features, on notice you control.

White-label work package

When you sell delivery under your own brand: we work under NDA, in your repositories and tooling, hand the IP through to your end client, stay off your client communication unless you bring us in under your lead, and sign off against acceptance criteria written before the build.

From commit to running in your cloud

A single delivery path you can read end to end: every change moves through the same gates, and the same path runs in reverse when something needs to be pulled back.

Commit
CI gates
Artifact
Staging
Production
Observability
Staging, Production and Observability run in your cloud accounts
Rollback path:ObservabilityArtifact or Production

Where uptime matters, we agree it as an SLA target, not a measured promise.

Website builder, agency template, or engineered build?

Not every site needs engineering. Here is where each option wins, so you spend on the one your case calls for.

Scroll to compare

Website builderAgency templateEngineered build
Fastest and cheapest to launch a simple page
Holds its Core Web Vitals as content and traffic grow
Accessibility you can stand behind, tested not assumed
Full control over SEO, structured data and AI-crawler output
Custom functionality, logins and real integrations
You own the code, not a subscription that holds it
Still maintainable and extensible years from now

When a builder or a clean template fits a simple marketing presence, we will tell you and point you to the lighter path instead of overbuilding.

Why Oronts

Why teams build the web with us

We are not the biggest agency you can hire. Here is why owners pick us for the web anyway.

You own everything

The code, the content model and the IP are yours, transferred on delivery. No builder subscription that holds your site, no theme licence, no vendor you cannot leave.

Senior and founder-led

The people who scope your site build it. Senior engineering on performance, accessibility and SEO throughout, no junior hand-off after the pitch.

AI-native, fast without cutting corners

A small senior team with an AI-assisted workflow ships quickly and keeps the frontend clean, typed and accessible, with performance budgets enforced rather than promised.

A fixed-price way to start

Our 90-day production pilot puts a first usable, measured slice of the site live in a fixed scope and price, so you judge us on shipped Core Web Vitals before committing further.

build
$next build|
compiled successfully
generating static pages (5/5)
route / prerendered as static

Responsive by design

Every layout adapts, from widescreen monitors to pocket-sized screens.

Desktop
Tablet
Mobile

SEO and GEO built into the architecture

Technical SEO and AI-crawler readiness engineered in, not bolted on afterwards.

SSR & SSG

Server-side rendering and static generation for fast first paint and clean crawlability.

Structured data

JSON-LD for organization, breadcrumb and FAQ markup, validated at build time.

Meta & Open Graph

Per-page, per-locale titles, descriptions, canonicals and social cards.

Sitemaps & hreflang

Generated XML sitemaps with hreflang and lastmod for every localized route.

The stack we build on

Built for procurement

The answers a buying committee checks, before you have to ask.

Code ownership
Your repositories and IP, transferred on delivery under a work-for-hire agreement.
Hosting
Your cloud, your region, your tenancy. We deploy into your accounts, not ours.
Data
Customer data stays in the infrastructure we agree on, and we do not use it to train models.
Documentation
Architecture decision records, runbooks and a handover your team can act on.
Support
An optional retainer after launch. You are never locked into it to keep running.
Continuity
Full ownership and documentation mean any senior team can continue the work.
Data processing
An AVV per Art. 28 GDPR with a TOM document, ready to sign before we touch production data.
Subprocessors
A documented subprocessor list; you approve any processor before it is used.
Security review
We complete your security questionnaire and provide the security evidence your procurement needs, such as insurance proof or a BSI CyberRisikoCheck attestation, where required.
Acceptance
Defined acceptance criteria per milestone, so sign-off is against a written standard, not opinion.

Who owns what

Whether you contract Oronts directly or work with us under a prime or MSP, every link in the delivery chain has one clear owner.

Responsibility ownership across the delivery chain
ResponsibilityOrontsPrime / MSPYouCloud / model provider
Build & result responsibilityOronts owns Build & result responsibility
Code & IP ownershipYou owns Code & IP ownership
Hosting & infrastructureYou owns Hosting & infrastructureCloud / model provider owns Hosting & infrastructure
Data processing (AVV / TOM)Oronts owns Data processing (AVV / TOM)You owns Data processing (AVV / TOM)
Security questionnaire / attestationOronts owns Security questionnaire / attestationPrime / MSP owns Security questionnaire / attestation
Acceptance sign-offYou owns Acceptance sign-off
Incident response (agreed SLA)Oronts owns Incident response (agreed SLA)Prime / MSP owns Incident response (agreed SLA)

When we are not the right choice

  • Brochure sites a site builder serves better; we engineer platforms, not five-pagers
  • Pixel-perfect design replication with no engineering scope behind it
  • Shortcut SEO; we build for rankings that survive algorithm updates

Engagement levels

Oronts works with serious teams that need senior delivery, not low-cost outsourcing.

Production Pilot
from 25k EUR
Custom software and AI projects
from 50k EUR
Ongoing technical retainers
from 15k EUR/month

Exact pricing depends on scope, responsibility, delivery speed, team size, integrations, support expectations and production risk.

Frequently Asked Questions

React with Next.js is our default recommendation for most projects. It has the largest ecosystem, best server-side rendering support, and strongest TypeScript integration of any frontend framework. The component model scales well from small marketing sites to large enterprise dashboards, and the developer talent pool is significantly larger than alternatives, which matters when your team needs to maintain the codebase long-term. We use Vue when it is a better fit for the team or project, particularly when your existing developers are more productive in Vue or when you have a Vue codebase that needs extension rather than replacement. Angular is another option we support, especially for large enterprise applications where its opinionated structure and built-in dependency injection provide consistency across big teams. Ultimately, the framework choice should match your team's skills, project requirements, and long-term maintenance plan rather than following industry trends.
We use WordPress as a headless CMS but build the frontend with React or Next.js for better performance, security, and design flexibility. In a headless setup, WordPress handles content management through its familiar admin interface while the frontend is a separate Next.js application that fetches content via the WordPress REST API or WPGraphQL. This gives your content team the editing experience they already know while delivering sub-second page loads, better SEO through server-side rendering, and a modern component-based frontend that designers and developers prefer working with. The headless approach also eliminates many WordPress security vulnerabilities since there is no publicly accessible PHP frontend to attack. If you need a traditional WordPress site with themes and plugins, we can recommend trusted partners who specialize in that approach. Our strength is in custom-built frontends where performance, accessibility, and design precision are priorities rather than out-of-the-box template solutions.
Accessibility is part of our development process, not a checklist applied at the end. We use semantic HTML elements like nav, main, article, and button instead of generic divs, which gives screen readers and assistive technologies the structural information they need to navigate content effectively. ARIA attributes are added where semantic HTML alone is insufficient, such as for custom dropdown menus, modals, and dynamic content regions. Every interactive element is keyboard-navigable, and we test tab order and focus management throughout development. We run automated accessibility scans with tools like axe-core in our CI pipeline to catch regressions before they reach production. Manual testing includes screen reader verification with VoiceOver and NVDA, keyboard-only navigation testing, and color contrast validation. Our target is WCAG 2.2 AA compliance at minimum. For clients in regulated industries like government or healthcare, we can target AAA compliance where required.
Yes. We audit existing applications for performance, accessibility, and code quality, then implement improvements incrementally without disrupting your users or development workflow. The audit typically takes 1-2 weeks and produces a prioritized report covering frontend performance metrics like Largest Contentful Paint and Cumulative Layout Shift, accessibility gaps measured against WCAG 2.2 standards, code architecture issues including duplicated logic and missing error handling, and security vulnerabilities in dependencies. From there, we work through the highest-impact items first. For example, if your application scores poorly on Core Web Vitals due to unoptimized images and render-blocking JavaScript, we address those issues first because the improvement is immediate and measurable. We can work alongside your existing team, submitting pull requests through your normal review process, or take ownership of specific areas while your team focuses on feature development. Each improvement is deployed independently so there is no big-bang release.
Every project is mobile-first by design. We start with the smallest viewport and progressively enhance the layout for larger screens rather than building for desktop and then trying to squeeze everything onto a phone. We test across real devices including various iPhone and Android models, tablets, and different browser engines like Safari, Chrome, and Firefox. Our responsive design approach uses CSS grid and flexbox layouts, fluid typography that scales smoothly between 320px and 4K displays, and responsive images with srcset and the picture element to serve appropriately sized assets for each device. We also test touch interactions, gesture handling, and viewport-specific behaviors like iOS Safari address bar resizing. Performance testing on mobile is critical because mobile networks are less reliable, so we optimize bundle sizes, implement code splitting, and use lazy loading to keep initial page loads fast on 3G and 4G connections.
We use server-side rendering or static generation with Next.js to ensure all pages are fully crawlable by search engines. Unlike traditional single-page applications where content is rendered only in the browser via JavaScript, our SSR approach sends fully rendered HTML to crawlers, which means Google, Bing, and other search engines can index your content immediately without executing JavaScript. For pages with content that rarely changes, we use static site generation at build time for maximum performance and cacheability. For dynamic pages, incremental static regeneration keeps content fresh while maintaining fast load times. Beyond rendering, we implement structured data markup using JSON-LD for rich search results, generate comprehensive XML sitemaps, configure canonical URLs to prevent duplicate content issues, and optimize meta tags for each page. We also set up proper Open Graph and Twitter Card markup so shared links display correctly across social platforms and messaging applications.

Show us the web platform

Bring the URL and the complaint. We answer with findings, not a pitch.