📝 AVUIX

We had a huge library live for a while, and from far away it looked impressive. Big number, lots of titles, clean UI. The problem was underneath that surface, too many of those games were not actually ready.

Some opened to a blank screen. Some had broken controls. Some depended on old Flash wrappers or external assets that were gone. A few technically loaded but were so laggy they were pointless. Keeping that live felt fake, and honestly it was embarrassing.

So we made a hard call and pulled the library. Not because we gave up. Because we care more about reliability than bragging rights. A game platform should not feel like a lottery where one click works and the next three fail.

Right now we're going title by title and testing everything manually. Load time, controls, audio, mobile behavior, keyboard support, and whether the game still works consistently over time. If a game fails, it gets cut. No exceptions.

We're also cleaning naming, organizing categories better, and removing filler titles that nobody actually wants. The relaunch list will be smaller at first, but every game in it should feel worth opening.

That process takes time, but it's the right call. We'd rather launch something tight and reliable than keep pretending a giant catalog is healthy when it isn't.

In the meantime, proxies are fully live and maintained. If you want to be notified when the library drops, hit us on the contact page and we'll keep you posted.

AVUIX is intentionally simple under the hood. Hosting is GitHub Pages with a custom domain, which gives us fast global delivery, free static hosting, and very little ops overhead. If you can serve files quickly, you can run a good browser platform.

The frontend is pure HTML, CSS, and JavaScript. No React, no build step, no bundler, no dependency maze. For a static-first product, that keeps deploys predictable and the site fast on low-power devices.

Auth is Supabase. We use email/password plus Google and GitHub OAuth because people want options, and Supabase gives us mature auth flows without forcing us into a giant backend stack just to support sign-in.

The proxy layer runs on Cloudflare Workers. Edge execution keeps latency low, especially for users far from a single region. It also gives us easier scaling and better resilience than self-hosting proxy logic on one box.

UI customization is all CSS variables plus localStorage. Accent colors, motion settings, compact mode, and font sizing are all client-side. It's lightweight, instant, and doesn't require accounts or server sync to feel personalized.

We register a service worker so the platform has PWA behavior and better repeat-load performance. It's not trying to pretend this is a native app — it's just the right web primitive for caching and reliability.

For observability, we run Sentry for error tracking and GA4 for analytics. That gives us enough signal to fix real issues fast without turning the product into a tracker-heavy mess. The whole stack decision comes down to three things: simplicity, speed, and minimal vendor lock-in.

Everybody says "instant play" now, but most platforms still do extra network hops before you see anything usable. If a click triggers three external services and a remote iframe dependency, that is not instant play.

For us, instant play means the game file itself is already hosted by us as static HTML, not fetched from a random third-party API at click time. If our CDN is up, the game should open. Simple as that.

Delivery matters too. GitHub Pages gives us distributed caching and Cloudflare helps at the edge for related services. The closer files are to users, the less time gets burned before first paint.

Game file weight is a real constraint. Massive bundles and unoptimized WebGL builds destroy load times on average hardware. We favor lean titles and optimized packages because performance targets have to include low-power devices, not just gaming PCs.

The manifest system is the quiet workhorse here. We keep structured metadata locally so the platform can render listings and links without waiting on third-party catalog APIs. Fewer moving parts means fewer cold-start delays.

This is also why we avoid the "embed whatever works today" approach. Some sites depend on external hosts being alive forever. When those hosts go down, the library breaks. Self-contained hosting is less flashy but way more reliable.

Our bar is under 2 seconds to first interaction on a Chromebook over average WiFi. We're not fully there yet across every title, which is exactly why the library is being rebuilt right now. Long term, WebGPU is the next big unlock for browser gaming performance, and we're watching that closely.

First priority is the game library relaunch, and it stays first until it's done right. We're testing every title, removing broken entries, and tightening quality standards before anything goes back live.

In parallel, proxy stability work is ongoing. Reliability beats feature hype. If people click a proxy endpoint, it should respond quickly and stay up. That's basic trust and we treat it like core infrastructure.

After that, favorites and bookmarks are next for logged-in users. The account system already exists, so this is about giving users practical tools to save and return to what they actually play.

Leaderboards are also planned using the current game-api.js score foundation. We're not forcing social features, but competitive stats done cleanly can make the platform more fun without adding noise.

Push notifications are on the list too. Service worker support is already in place, so the next step is opt-in alerts for meaningful updates, not spam blasts.

We still plan to expand the library toward 2,000+ titles over time, but only if quality stays high. Growth is not the goal by itself. Reliable play is.

AdSense is applied and pending. If approved, we'll use non-intrusive banner placements only — never interstitials, never autoplay video. Not happening: paid tiers, real-identity requirements, or social-network gimmicks. The rule is simple: every shipped feature must make AVUIX faster, more private, or more fun.