Thoughts on building Ziphay, the future of file compression, and what's coming next.
From a side project solving a personal frustration to a market-ready startup — the full story behind Ziphay, what makes it different, and where it's going.
Every developer and creator has been there. You have a 4 MB image, a client needs it under 1 MB, and every compression website you open is a disaster: fake "Download" buttons, ads that hijack your screen, limited format support, and tools that store your private files on random servers for days.
I was frustrated enough to build my own. What started as a single HTML file with a Canvas API call became something much bigger — and much more interesting.
Ziphay is a 100% browser-based file compression and upscaling platform. No servers, no uploads, no accounts required for core features. Every computation happens inside the user's own browser using:
The UI was built entirely from scratch — no CSS frameworks, no component libraries. Just a design philosophy borrowed from the things I love: clean, fast interfaces that feel inevitable once you're using them.
Ziphay's smart features — goal-based compression profiles (Web, Email, Social, Archive), automatic format selection, estimated size previews, and intelligent quality tuning — are all driven by logic that mimics what a human expert would do when deciding how to compress a file.
The roadmap includes integrating actual neural compression models (like those used in modern codec research) to run via WebAssembly directly in the browser. This would unlock genuinely AI-driven quality reconstruction — the same technology behind modern image codecs like AVIF and JPEG XL.
The global image optimization software market is growing steadily as web performance becomes a critical metric for businesses. Yet the current tools are either too technical (command-line tools like ImageMagick), too cluttered (ad-filled web apps), or too expensive (enterprise CDN solutions).
Ziphay targets the gap in the middle: prosumer creators, small development teams, and digital agencies who need reliable, private, fast compression without the enterprise price tag or the ad-filled experience.
Browser-based compression, upscaling, goal modes, before/after slider, dark/light mode, batch processing, format conversion.
User accounts via Firebase, compression history dashboard, cloud storage integration (Google Drive / Dropbox), Stripe-based Pro subscriptions, API access for developers.
ONNX-based neural compression models running via WebAssembly. Genuine perceptual quality reconstruction — not just resizing. AI-powered video compression using FFmpeg.wasm.
White-label Ziphay for agencies, REST API + webhooks for developer integrations, enterprise SLA contracts, team workspaces, custom branding.
Ziphay follows a classic freemium SaaS model. The free tier is genuinely useful and always will be — that's non-negotiable. Revenue comes from Pro ($9/month) and Enterprise (custom) tiers that unlock higher file size limits, batch processing, API access, and priority processing queues.
Three things that I believe no current competitor does all at once:
Building Ziphay has taught me more about product thinking, performance engineering, and user experience than any course I've taken. Every decision — from the teal colour palette to the multi-pass upscaling algorithm — was deliberate.
I'm building this as a solo developer while finishing my engineering degree. Every line of code is mine. I believe that's a feature, not a limitation — it means every feature I ship is something I personally tested, iterated on, and believe in.
Ziphay is just getting started. Watch this space.
The technical deep-dive into building a multimodal AI tutor with Gemini 2.5, Firebase, and a real-time HTML5 whiteboard.
Everything I learned building Ziphay's compression engine — from WebP quality curves to multi-pass upscaling.
The unexpected connection between filmmaking, CineAbstra, and building great software products.