The Developer

Why I built PixelFix Pro

A story about being genuinely fed up, and doing something about it.

It started with a passport photo. I needed to resize an image to exact ICAO dimensions for a visa application. Shouldn't take more than a minute.

The first tool I tried asked me to create an account before downloading my resized photo. The second uploaded my file to a server and showed me ads for 30 seconds while it "processed." The third gave me the correctly-sized photo but put a giant watermark across it and asked me to pay to remove it.

This is a passport photo. My face. Being uploaded to servers I know nothing about, by companies I've never heard of, for a task that requires zero server-side processing.

The reason these tools upload your files isn't technical necessity — it's a business model. If your file passes through their server, they can log it, analyse it, gate the download behind a paywall, and force account creation. A browser has had everything needed to process images locally since HTML5 Canvas became standard. The technology was already there. Someone just needed to build it honestly.

So I built it myself

PixelFix Pro runs 100% in your browser. Every operation — resize, compress, crop, convert, filter, watermark, PDF tools, batch editing — happens using your own device's GPU through browser-native Canvas APIs. The file goes straight from your storage to your downloads folder. My servers are never involved, because I don't have image-processing servers.

The way I prove this: load the site, disconnect your WiFi, then use any tool. Everything still works. That's not possible if anything was being sent anywhere.

What's different about how this is built

The site uses Astro — a framework that outputs pure static HTML/CSS/JS with no server runtime. It's deployed to Cloudflare's edge network. There's no backend, no database, no user accounts, and no image-processing pipeline — because there doesn't need to be one.

Image processing uses OffscreenCanvas, which runs on the browser's GPU thread rather than the main thread, keeping the UI responsive even while crunching large images. PDF tools use pdf-lib, jsPDF, and PDF.js — mature open-source libraries that execute entirely client-side. HEIC conversion uses heic2any, also client-side. No pixel of your image ever touches my servers because my architecture doesn't include any.

Honest about the trade-offs

I run ads to keep this free. Google AdSense displays standard banner ads. That's how the site pays for itself. The ads are served by Google — they use cookies, which is disclosed in the privacy policy.

What I don't do: sell your image data (I can't access it), require an account, gate features, add watermarks, or run ads so aggressive they make the tool unusable. I removed Monetag's vignette and in-page push formats deliberately — a pop-under ad next to "your data never leaves your device" would be incoherent.

The site is built to last without maintenance. Tools work offline. No server to go down, no subscription to cancel, no database to maintain. Once deployed, it should keep working for years.

What's coming next

An EXIF metadata viewer and remover — a tool that shows you what your photos are secretly broadcasting (GPS coordinates exact to the metre, device model, timestamp, camera settings) and strips it clean. This is the privacy tool I keep wanting most, and I haven't found a good client-side version that doesn't require an upload.

Beyond that: a steganography tool (hide a message inside a photo, pixel by pixel, then decode it), and potentially more PDF tools. All following the same principle: if the browser can do it, the server shouldn't.

All the tools, in one place

☕ Support