Today, I’d like to share the story behind Accesserty, a small product ecosystem I built in about a month, aimed at improving the accessible web experience. From the moment users search to the moment developers build, I wanted to offer a calm, practical set of tools to make accessibility visible, understandable, and actionable.
Origin: Curiosity about Chrome Extensions
Back in 2023, I created my first Chrome extension — Report Website Issues — simply out of curiosity.
Now in 2025, with the rise of large language models (LLMs), I began thinking about how AI might help me build tools that address real-world accessibility challenges.
I experimented with assigning LLMs different “roles” — product manager, frontend dev, accessibility analyst — and used them to speed up system design and implementation.
Identifying the Problems
1. After searching, users still don’t know which link to click
Search engines consider SEO, performance, and content, but that doesn’t mean the top results are easy to use — especially for people who rely on keyboard navigation or screen readers.
Sites may appear technically sound but be visually cluttered, hard to navigate, or inaccessible by design. So I started wondering:
What if I could see how accessible a site is — right from the search result list?
It would save time, reduce frustration, and help everyone make more informed choices.
2. Users often have no way to report issues
Even if a user encounters a serious accessibility problem, it’s often unclear how to report it — or if it will even be taken seriously.
Worse, communication gaps happen. Like in a case I mentioned in an earlier post:
A user says: “I can’t browse your site using the keyboard.”
A support rep replies: “It works fine for me — maybe try again?”
But they’re talking about different things. One refers to keyboard focus logic; the other is just pressing the down arrow key. They both get frustrated.
So I realized there needs to be a “bridge” — someone or something that helps users explain issues clearly and helps developers receive usable feedback. Right now, LLMs might help with summarization, but the cost and accuracy are still limiting.
3. Is accessibility really expensive for developers? With recent EU regulations, many UI frameworks now include more accessible components — that’s progress. But passing machine audits isn’t enough.
Compliance ≠ usability.
Take text, for instance. A machine can check if it’s there — but only a human can tell if it actually describes the image meaningfully in context.alt
Even if LLMs could generate decent descriptions, that still involves API costs, image parsing, and prompt design — not to mention hallucination risks.
That led me to a fundamental insight:
The earlier you address accessibility, the lower the cost.
From planning to design to engineering, everyone needs to be aligned early. This diagram sums it up well:
Solutions: Simulate, Detect, Report
Based on these observations, I created a journey map to visualize the pain points across users, developers, and site owners.