INVISENT®

04 / Capabilities

Development

Production engineering: typed content, server rendering, performance budgets and a build someone else can maintain.

Part of03 Build

What it is

Creative development is the engineering half of digital work: turning a design system into software that renders fast, stays accessible, survives content changes and can be handed to another team without archaeology.

The mechanism

Rendered on the server, measured in the browser.

A typed content model feeds server rendering; what can be static is static; the browser receives only the script that has a job. Budgets are agreed before the build and integrations run behind validation, never in front of it.

development / architectureCoordinated
01System architecture

From content to the screen

Server

  • Content modeltyped contentStructured
  • Renderingserver componentsVerified
  • Static outputprerenderReady

Delivery

  • Edgecache · headersMonitored
  • Browseronly what has a jobActive
02Performance

Budgets set before the build

  • Largest paintLCPMonitored
  • Script weightscript budgetVerified
  • Layout stabilityCLSVerified
  • Fontssubset · self-hostedReady
  • Motion costtransform · opacityVerified
03Integration layers

Validation first, then everything else

  1. webhookInputFrom a person or a systemTracked
  2. zodValidationRejected before it costs anythingVerified
  3. route handlerHandlingOne place per operationStructured
  4. retry · incident idDeliveryFailures are named, not swallowedMonitored
  5. system of recordRecordOne source of truthOwned
A drawing of the structure this discipline produces. It describes the shape of the work, not a product, and shows no client data.

Who it is for

  • Companies whose current site is slow, fragile or impossible to edit safely
  • Teams migrating off a page builder that has outgrown its purpose
  • Organisations that need the site to integrate with systems they already run

What usually breaks

  • A marketing page ships three megabytes of JavaScript to display text.
  • Content changes require a developer, so the site goes stale between releases.
  • Nothing is typed, so a missing field becomes a production error instead of a build error.

What we change

01

Server first, client only where it earns it

Pages render on the server. Client-side JavaScript is added for the few parts that genuinely need the browser, and nowhere else.

02

Content is a typed model

Every content type is declared once. The compiler catches a missing field before a visitor does, and the same model feeds the page, the metadata and the structured data.

03

Budgets, not intentions

Route-level transfer budgets are set at the start and checked before release. Performance is a constraint we design inside, not a score we chase afterwards.

Deliverables

  • Production application with typed content models and a documented architecture
  • Component implementation matching the design system, token for token
  • Rendering strategy, caching and image pipeline
  • SEO, structured data and sitemap infrastructure
  • Security headers, form handling and validation on the server
  • Handover documentation: how to add a page, a project, an article or a language

What it does not do

  • We do not take over a codebase we have not read. An audit comes first, and it sometimes concludes that a rebuild is the cheaper option.
  • We do not build features that will have no owner on your side after launch.
  • Performance targets are agreed as budgets before the build. We do not report a lab score as an outcome.

Process

01

Architecture

Stack, rendering model, content model and integration boundaries, written down before the first commit.

02

System implementation

Tokens and components first, pages second. The same order the design took.

03

Instrumentation

Analytics, consent, error handling and structured data wired as part of the build, not bolted on.

04

Release

Lint, type check, tests, build, accessibility and performance review, then deploy with a rollback path.

Questions

What do you build with?

Next.js with the App Router, TypeScript in strict mode, and a token-driven CSS architecture. We choose boring, well-supported technology for client work: the goal is a codebase another team can still maintain in three years.

Can we edit content ourselves?

Yes. We either connect a headless CMS or ship typed content files with a documented editing flow, depending on how often content changes and who changes it. The content model is designed so switching later does not mean rebuilding.

Who owns the code?

You do. The repository, the pipeline and the documentation are yours at the end of the engagement, with no licence tied to us.

Can you take over a codebase somebody else wrote?

Sometimes. We read it first and say plainly whether it is worth continuing or worth replacing. Those are two different bills, and we would rather tell you before the work starts than in its third month.

Start

Have a system worth building?

Tell us what the business needs to do. We will tell you what it takes to build it, and what it takes to keep it running.