What is a browser-based AI image colorization tool?

It's DeOldify running via ONNX in your browser—no servers, keeps images private, works offline after load.

How do I build my own browser AI colorizer?

Grab the code: load ONNX Runtime, fetch quantized model, preprocess to 256x256 tensor, infer, postprocess to canvas. Service worker caches.

Does browser AI colorization work on mobile?

Yes, but initial 25MB load needs WiFi; iOS shines, Android improves with WebGPU.

🤖 AI Dev Tools

Your Grandma's Faded Photo Just Got Color—Entirely in the Browser, No Servers Required

Click. Wait. Color blooms across that black-and-white family portrait—all without your photo ever leaving your laptop. Browser-based AI image colorization is here, and it's smarter than server hype.

theAIcatchup Apr 09, 2026 3 min read

Before-and-after of black-and-white photo colorized by browser AI tool

⚡ Key Takeaways

Client-side AI slashes cloud costs and boosts privacy—no photo uploads ever. 𝕏
25MB DeOldify ONNX model runs offline after one cache, with sub-second repeats. 𝕏
This mirrors WebGL's browser shift; expect 40% more client ML by 2025. 𝕏

Published by

theAIcatchup

Ship faster. Build smarter.

#DeOldify #ONNX Runtime #ONNX Runtime Web #WebAssembly ML #WebGPU #browser AI #image colorization #offline AI

Worth sharing?

Get the best Developer Tools stories of the week in your inbox — no noise, no spam.

Originally reported by dev.to

⚡ Key Takeaways

The 60-Second TL;DR

theAIcatchup

Share this article

Worth sharing?

Related Stories

WebGPU Unlocks Client-Side AI Background Magic

npm install TinyTTS: Offline TTS That Fits in Your Pocket — And Crushes Cloud Dependencies

Browser AI Upscalers: Ditch Servers, Keep Photos Private Forever

Browser OCR Revolution: Tesseract and PP-OCRv5 Clash Client-Side

Stay in the loop