I’m a developer. I don’t usually post in QA communities, so apologies if I’m not in the right place — but I think what I built might be relevant to some of you.
For years, I’ve been struggling with visual regression testing on my projects. I tried Selenium-based approaches, but I was spending more time writing and maintaining tests than actually building
features. At some point, we even tried having someone on the team do manual visual checks — but when testing isn’t your job, it’s a thankless task and regressions still slip through.
Every solution I found was built for developers. And it hit me: the person who actually knows what should or shouldn’t change visually is usually the QA professional, not the dev.
So I built my own tool. The idea is simple:
You record your scenarios — no code, no scripts to maintain
It captures and historizes every version — you can see how your pages evolve over time and compare any version with another
It detects visual regressions automatically — pixel-by-pixel comparison
You decide what’s a real change and what’s not — the tool doesn’t make the call, you do
I showed it to a QA Engineer I work with. His reaction was: “This is a tool for me.” That one sentence changed how I saw the project.
It’s called Delta-QA and it’s a first version — rough around the edges, I know. I built it for myself and I’ll keep using it regardless. But if it can help QA professionals, that’s even better. I’m
also considering open-sourcing it if there’s enough interest.
I’d genuinely love your feedback:
Is visual regression testing a real pain point in your workflow?
Does this approach make sense to you?
What would you need to actually use something like this?
Hey, welcome to the community. It’s good to have you hear, and definitely a good space to ask for this kind of feedback - I’m glad you found us.
Is visual regression testing a real pain point in your workflow?
It’s something I want to introduce on my project, but for .pdfs - if there would be a good solution for that, it would be great
Does this approach make sense to you?
I usually don’t jump towards no-code tools, because I like to have more control over what it does, and flexibility to suit my needs. I get what you’re saying with the pixel-perfect approach, but things can render differently under different configurations, so this could create a lot of noise. So the control element is there, but not in the way I’d personally look for it. If you could combine a no-code required but possible approach with some sort of “smart detection” of meaningful changes (perhaps thresholds which could be set), I think that would be going in the right direction
What would you need to actually use something like this?
I’m curious about things like where data is hosted, privacy, costs, limitations. If it were a self-hosted, fully open source solution (that worked with .pdfs) without snapshot / check limits, it’s definitely something I would evaluate against other options
Delta-QA is designed to be completely free and runs locally on desktop, so everything stays on your machine — no external hosting, no data leaving your environment.
If you want to get a quick feel for the visual comparison approach, I’ve put a simple tool online ( Delta-QA website ) that lets you compare two HTML pages side by side.
Would be really interested to hear your thoughts after trying it.
There are significant scrolling bugs on your website. The overlapping scrolling front page looks slick but depending on where your mouse cursor is on the page, can result in being unable to scroll down at times (to reproduce, try scrolling up a few slides, then try scrolling down.)
More egregious, if you scroll all the way to the bottom (footer in view) after a brief delay the page will auto scroll up slightly taking the footer out of view again.
These problems on a website advertising a product which is supposed to be a QA solution would turn off a lot of potential users. It’s the first thing they will see of your product – how can it help them if it hasn’t helped itself?
I haven’t tried the software itself. I might give it a go on my personal machine which I don’t have access to currently.
Thank you very much for the detailed feedback. We’ve taken these remarks into account internally.
We realized that the scrolling experience was not really suitable on desktop with a mouse, especially because the behavior differs a lot between a trackpad and a traditional mouse wheel. The effect could feel smooth in some situations, but also created frustrating behaviors like the ones you described.
As a result, we decided to significantly simplify the website in favor of something more stable.
You are absolutely right that these kinds of details matter even more for a QA-oriented product — the credibility of the product also comes from the experience of the website itself.
Thanks again for taking the time to share this feedback. We’re now looking forward to hearing your thoughts on the application itself once you get a chance to try it.