Why we built automated screen reader testing
Our origins are in digital marketing, handling ADA compliance requirements and responding to accessibility lawsuits for clients. This is what we learned there — and why monitoring is free, scanning is unmetered on every paid plan, and journey testing is what you pay for.
Our story
Where this started
Client sites, client deadlines, and — increasingly — client ADA problems. Thorough accessibility testing is laborious and complex, and nobody can honestly guarantee the result. We spent years helping clients meet those requirements, and a good deal longer helping them respond once a letter had already arrived.
Then the letters stopped being unusual
In some industries they became close to unavoidable. Sometimes because a site genuinely locked people out. Sometimes because a firm had found an efficient route to a quick settlement. From where our clients sat, the difference was academic — the letter arrived either way.
What actually held up
The strongest position we ever saw a client in was one built on evidence: consistent testing, monitoring and fixing, documented over time. When a gap surfaced that the testing had missed — and gaps surface — the record still showed a team taking the standard seriously and watching its own adherence to it.
That is not a legal guarantee, and we do not sell one. It is the difference between having an answer and not having one.
Most of it was already solved
Rule-level validation has been automated for years, and much of it is free. axe-core catches a large class of failures, and a developer who knows what they are doing can run those scans and fix the markup long before it reaches production. For that part, you do not need a vendor at all.
Except for one thing
Already solved
Rule scanning
Automated for years by open tooling. Catches a real and useful class of failures, and costs nothing to run.
$0
with open tools like axe-core, and unmetered on every paid Accessible plan
Not solved
Screen-reader journeys
Whether a blind user can finish sign-in, search or checkout, end to end. No tool ran it. It was manual work, quoted per engagement.
$5–10k
a round, in the quotes our clients saw
At that price it happened once, or it did not happen. Usually it did not happen — and the one category of barrier that actually stops a person from completing a task was the category nobody was checking.
What changed
AI made the journey repeatable. Accessible drives a real browser using the navigation patterns of a screen-reader user — headings, landmarks, roles, tab order — through the entire task, and reports the step it broke on and what was announced there. Not once, as a contract. Every release, if that is what you want.
Monitoring your real visitors is free. Scans come unmetered with every paid plan, because a scan is where accessibility starts, not where it ends. What you pay for is the journey testing.
What we hold ourselves to
Three rules we wrote down before we needed them, and enforce in the product itself.
We never assert more than we measured
A figure we did not measure does not get rendered. Missing data shows as “not evaluated”, never as a zero — because on a compliance product a fabricated number is exposure, and the absence of data is itself honest information.
A clean scan is never sold as proof
Passing every automated rule does not mean a screen-reader user can finish your checkout. We will not present one as evidence of the other, which is the single failure this whole product exists to answer.
Our own interface holds to the standard
WCAG 2.2 Level AA is the floor for Accessible itself, not just for what we test. Shipping an inaccessible accessibility product would argue the other side’s case.
Start with the part that costs nothing.
Monitor one domain free, with no card and no expiry, and see which barriers your real visitors meet. Check any single page free, no account needed. When you want every URL scanned and a journey run end to end, a plan adds both, starting with a 7-day trial.