Custom / Static

Help with your custom or static site.

When the off-the-shelf platforms don’t fit — when you need a true performance ceiling, full content ownership, an API-driven architecture, or a frontend that does not look or behave like everyone else’s — custom is the right answer. We design and build production JAMstack and headless sites, and we keep them shipping after launch.

The custom toolkit, in concert · Static Site Generators · Astro · Next.js · Hugo · JAMstack · Performance · Edge Hosting · Custom Builds
01 — What’s Included

Six capabilities.
One coherent stack.

Aureole builds production custom sites where the project genuinely needs them. Most clients are better served by WordPress or a hosted platform — but for projects with specific performance, customisation, content-architecture, or operational requirements that hosted platforms cannot meet, custom is the right call.

The work spans framework selection through deployment, ongoing operations, and the SEO discipline that keeps a JavaScript-rendered site findable. We will tell you honestly when custom is the right tool — and when it is not.

N° 01

Static-Site Builds with Modern Frameworks

Foundation

We build production sites with the frameworks that actually fit the project — not whatever is trending. Astro for content-heavy marketing sites where minimal JavaScript and excellent SEO are the priority. Next.js or Remix when the project needs server-side rendering, dynamic data, and React’s component ecosystem. Hugo when raw build speed and simplicity matter more than developer ergonomics. Eleventy when the team prefers a smaller pipeline. SvelteKit for projects that prioritise bundle size and runtime performance. Framework choice is a decision with five-year consequences — we treat it accordingly.

N° 02

Headless CMS Integration

Editorial

For sites that need editorial workflows but do not want a traditional CMS rendering the frontend, we integrate headless content systems. Sanity for projects with rich structured content and complex editorial workflows. Contentful for enterprise-style content operations. Payload CMS for self-hosted ownership with developer-friendly APIs. Headless WordPress when the editorial team already knows WordPress and the engineering team wants a custom frontend. Schema decisions in a headless CMS are even harder to reverse than in a traditional CMS — we get them right at the start, not after the first launch.

N° 03

API-Driven Architectures

Data layer

Custom sites often pull data from internal APIs, third-party services, e-commerce backends, search providers, or analytics platforms. We design the data layer carefully: which data is fetched at build time versus request time versus client-side, how caching layers interact, how stale-while-revalidate patterns keep the site fast, how authentication flows for protected content work, and how the site degrades gracefully when an upstream service is slow or down. A sloppy API integration is one of the most common ways a custom site ends up slower than the WordPress site it replaced — we do not let that happen.

N° 04

Performance Optimisation for Modern Frameworks

Speed discipline

Custom static sites are not automatically fast. They have the potential to be very fast, but only when built with discipline. We optimise the controllable factors aggressively — bundle size auditing, dependency replacement, code-splitting, image optimisation with proper format selection and responsive loading, font loading strategy with preload and display: swap, and third-party script management. Lighthouse scores in the high 90s are achievable but not automatic — they come from disciplined engineering, not framework defaults.

N° 05

SEO for SPAs and JavaScript-Rendered Sites

Findability

Custom React, Vue, or SvelteKit applications can absolutely rank in search — but only if they are built with SEO in mind from the start. We work the technical SEO checklist for JavaScript-driven sites: server-side rendering or static pre-rendering for content that needs to be crawlable, proper meta tag handling per page, sitemap generation that reflects the actual indexable pages, robots.txt configuration, structured data implementation, canonical management, and hreflang for international sites. We validate with Google’s URL Inspection and rendered-DOM checks — not by trusting that “the framework does SEO.” See our full SEO discipline →

N° 06

Hosting & Deployment Operations

Operations

Custom sites need real deployment infrastructure. We deploy on Vercel, Netlify, or Cloudflare Pages for projects that fit those platforms’ models. We deploy on dedicated VPS or container infrastructure for projects that need more control, or that have data-residency requirements. We set up CI/CD pipelines that test before deploying, preview environments for every pull request, monitoring and alerting for the production site, and rollback procedures for when a deployment goes wrong. The operational layer is half the work on a custom site — and it is the half that gets neglected most often.

02 — Common Problems

The four ways
custom sites break.

Custom builds are powerful when the requirements call for them — and brittle when they don’t. Most of the custom-site work that walks through our door is not greenfield. It is rescue: an existing build that has drifted into one of these four failure modes.

i

SEO invisibility for JavaScript-rendered content

The most common custom-site failure mode. A React or Vue app renders beautifully in the browser, but search engines see an empty HTML shell because the content is loaded client-side after the initial response. The result is content that exists but does not rank. The fix is server-side rendering or static pre-rendering of all indexable content — implemented properly, with the SEO meta data rendered server-side too, not just the body content.

ii

Content management without a real CMS

Custom sites built without a CMS work fine for engineering teams that are comfortable editing markdown in a Git repository. They do not work for marketing or content teams who need to update copy, swap images, schedule posts, or manage editorial workflows without engineering involvement for every change. We retrofit headless CMS integrations, designing schemas to match the existing content model and building editor experiences that are genuinely usable.

iii

Deployment complexity & operational fragility

Custom sites often launch with deployment processes that worked when one engineer was building the site, and that fall apart when the team grows or that engineer leaves. The site is live but no one knows how to update it confidently. We rebuild the operational layer: documented deployment procedures, proper environment management, CI/CD that runs tests before deploying, preview environments, monitoring with real alerting, and runbook documentation.

iv

Scaling issues at real traffic

A custom site that works fine at low traffic can fall over at higher traffic in surprising ways — a server-side rendering bottleneck on a serverless platform with cold-start problems, an upstream API that cannot handle the load, a database query that worked on test data but not on real data, a CDN configuration caching the wrong things or nothing at all. We diagnose with the actual production traffic patterns and fix the specific bottleneck.

03 — Who It’s For

When custom is
genuinely the right call.

For most marketing sites, WordPress with a clean theme is faster to build, easier to maintain, and just as effective. We recommend custom only when the project has a real reason for it — and these are the four shapes that reason most often takes.

Custom only earns its complexity when the project has specific requirements that hosted platforms genuinely cannot meet.

  • i Marketing sites with a true performance ceilingSub-second TTFB, near-zero JavaScript on content pages, edge-rendered globally. WordPress with optimisation can get fast — custom static can get faster, and the difference matters when conversion or ranking is on the line.
  • ii Custom apps with unique requirementsReal-time features, custom search, complex interactive applications, calculators, configurators, dashboards. Functionality that exceeds what plugin-based platforms can deliver cleanly.
  • iii Complex content models that benefit from a headless architectureHighly relational data, multi-channel publishing, structured content reused across many surfaces. The kind of editorial operation where the schema is the product.
  • iv Specific compliance, scaling, or integration needsData-residency requirements, custom authentication flows, integrations with internal systems, or operational constraints that hosted platforms cannot accommodate. The decision is dictated by the requirements, not the trend.
  • v Engineering teams that want full ownershipGit-based workflows, code review on every change, full ownership of the codebase, the ability to ship custom features without waiting on a vendor’s roadmap. When the team genuinely wants this — and has the discipline to operate it — custom is the honest answer.

If none of these shapes describe your project, custom is probably wrong for you — and we will say so. The fastest, cheapest, most maintainable answer for the majority of marketing sites is still a well-built WordPress site with proper SEO. Honesty about platform fit is part of the service.

04 — A complimentary report

Wondering why your site feels slow?

Send us your URL. We’ll send back a Premium Performance Report within 48 hours — page speed, Core Web Vitals, accessibility, and a prioritised fix list ranked by impact on rankings and conversion. For custom sites we will identify the bundle-size, dependency, and rendering bottlenecks that are usually the actual cause.

No sales call required.

Custom is not a style choice. It is a performance commitment — and the discipline to honour it every day after launch.
— The Aureole Practice —
05 — Frequently Asked

Questions we get
about going custom.

If a question is missing here, the contact link at the foot of the page goes straight to the person who would answer it. No ticket queues, no funnels.

i When is a custom site the right choice instead of WordPress or a hosted platform?
When you have specific requirements that hosted platforms cannot meet. The most common reasons we recommend custom: performance ceilings that WordPress with optimisation cannot reach (sub-second TTFB, near-zero JavaScript on content pages, edge-rendered globally); content models that are genuinely complex and benefit from a headless architecture (highly relational data, multi-channel publishing, structured content reused across many surfaces); engineering team preferences and ownership requirements; or specific functionality (real-time features, custom search, complex interactive applications) that exceeds what plugin-based platforms can deliver cleanly. For most marketing sites, WordPress is the better answer — we recommend custom when there is a real reason for it.
ii How do you handle SEO on a Next.js or React app?
With server-side rendering or static generation as the foundation, not as an afterthought. Every page that needs to be indexed must produce a complete HTML response with the right meta tags, structured data, and content body before any client-side JavaScript runs. Next.js App Router with proper RSC usage, or static export for sites that fit that model, gives us this by default. We then implement the rest of the technical SEO stack: per-page meta data including Open Graph and Twitter Card, structured data per page type, sitemap generation that reflects the actual indexable routes, robots.txt configuration, canonical tag management, and hreflang for international sites. We validate with Google’s URL Inspection tool and rendered-DOM checks — not just by trusting that “Next.js does SEO.”
iii Which static-site framework do you recommend?
It depends on the project. Astro is our default for content-heavy marketing sites that want the SEO and performance benefits of static rendering with the option to use React, Vue, or Svelte components where they help. Next.js is right when the project needs serious dynamic capability, when the team is React-fluent, or when the deployment platform is Vercel and the integration matters. Hugo is right when build speed dominates the requirements and the content model is simple. Eleventy is right when simplicity is the priority and the team prefers a smaller toolchain. We pick based on the project, not on framework loyalty.
iv What about headless WordPress?
Real and useful for specific cases. Headless WordPress means using WordPress purely as the CMS — content editing happens in the familiar WordPress admin — while the public site is rendered by a custom frontend (typically Next.js or Astro) that fetches content from WordPress via REST or GraphQL. This gives the editorial team a CMS they already know, and the engineering team a custom frontend with the performance and flexibility advantages of a static or server-rendered architecture. The trade-off is operational complexity — you are running and maintaining two systems instead of one. We recommend headless WordPress when the editorial team specifically needs WordPress-grade authoring and the frontend specifically needs to be custom. For most projects, full WordPress with a clean theme is simpler and just as effective.
v How do you handle hosting for custom sites?
Match the hosting to the architecture. Vercel is the default for Next.js projects — the integration is tight and the platform is purpose-built for the model. Netlify and Cloudflare Pages work well for Astro, Hugo, Eleventy, and other static-first frameworks. For projects with specific data-residency requirements (Canadian data must stay in Canada, etc.), or projects that need more control than serverless provides, we deploy on dedicated infrastructure (DigitalOcean, AWS, Hetzner) with proper container orchestration. Choice of hosting is a real decision with real cost and operational implications — we make it deliberately, not by defaulting to whatever the framework’s documentation recommends. Hosting and operations sit alongside our broader IT Services work.
vi Can you take over a custom site that another agency or team built?
Yes, regularly. Most of our custom-site work is taking over existing builds and improving them — fixing performance, adding SSR for SEO, restructuring messy CMS integrations, improving the deployment pipeline, retrofitting monitoring, or just bringing the site under a maintenance plan that keeps it shipping reliably. We start with a code audit and an operational audit: what is the codebase, how is it deployed, what is the test coverage, what is the documentation state, what are the known issues, and what is the highest-leverage place to start. Then we work through the issues in priority order, transparently, with clear scope and reporting.
vii Will my Lighthouse scores actually be in the high 90s?
On the projects we build from scratch and operate ourselves, yes — that is the bar we hold ourselves to, and it is achievable on a static or properly server-rendered site with disciplined engineering. On rescue projects, the honest answer is “eventually, in priority order.” A site that scores in the 40s today might land in the 70s after one round of fixes, the 80s after the second, and the 90s once we have rebuilt the dependency budget, image pipeline, and font loading strategy. We will not promise a specific number on day one — we will publish the baseline, the trajectory, and the prioritised fix list, and we will report honestly on where we are against it.
viii Do you do greenfield custom builds, or only rescues?
Both. Greenfield builds are where the framework, content model, and operational layer all get designed together, and the discipline shows up in the first commit rather than the fiftieth. Rescues are where we take an existing custom site that has drifted and bring it back to a healthy operating state. Greenfield is faster and cleaner for the team that ships it; rescue work is where the highest-leverage improvements live. We are happy to do either — but in both cases we start with the same question: does this project actually need to be custom? If the answer is no, we will say so before we quote the work.
The Invitation

Ready to make your
custom site work harder?

Send us a message and we’ll write back within one business day. No automated funnels, no follow-up calls until you ask. Or start with a free Performance Report — same team, same care, no obligation.

Mon–Fri · 9–6 PT support@aureoleintelligence.com Reply within 1 business day