Every page captured. Every layer read. Every criterion decided.
The EnableUser platform crawls your whole estate, walks the signed-in journeys your customers depend on, captures each screen at every layer, and takes all 55 WCAG 2.2 Level A and AA criteria to a decision, with the evidence behind every finding.
- 25,000+web pages audited
- 3,000+app screens and states audited
- 90,000+pages inventoried
- 500+regulatory filings delivered
The difference
What a scan sees, and what the platform sees.
Most automated testing tools can judge only part of WCAG, often as little as 10–20% of its criteria, and only on the pages they are pointed at. Nothing in scope is skipped here: we do not audit one page in fifty or stop at a tool check.
| A typical automated scan | The EnableUser platform | |
|---|---|---|
| Pages covered | The pages it is pointed at, often a sample | Every page in scope, mapped by the crawler and grouped by template |
| Signed-in journeys | Stops at the log-in screen | Log-in, two-factor, KYC, account opening and order entry, walked in your own session |
| What is read | The page's code | Rendered pixels, the code, the accessibility tree and, on apps, what the screen reader actually says |
| Error states | Not reached | Forms submitted in a sandbox to bring up every error, with nothing sent to your servers |
| WCAG criteria decided | Part of WCAG, often 10–20% | All 55 Level A and AA criteria, each with a stated reason |
| Findings | Rule hits, many of them false | One row per screen, element and criterion, each surviving an adversarial review |
| Re-testing | Run the scan again | Every past finding matched to the current screen and re-measured on the same element |
Who it is for
The people who depend on every screen working.
A blind investor reading a contract note on a braille display. A grandfather opening his first demat account with a grandson beside him. The platform tests what they meet, screen by screen, so neither needs help to finish what they started. Press play to hear a website the way a screen-reader user does.
People who are blind or have low vision
Screen readers, braille displays and magnifiers. Every control needs a name, every image a meaning, and every page an order that makes sense when heard.
People who use a keyboard or switch
No mouse, no touch. Every action must be reachable by Tab, Enter and the arrow keys, with focus that can always be seen.
Older customers
Larger text, steadier pace and clear steps. Time limits that can be extended, and buttons large enough to press.
People who are deaf or hard of hearing
Captions on every video, transcripts for audio, and alerts that are shown, not only sounded.
People with cognitive or learning disabilities
Plain language, forms that explain their errors, and journeys that do not punish a pause.
The teams who answer for it
Compliance officers, product owners and developers who need the barriers gone and the proof on file for regulators, auditors and buyers.
How the platform works
Six stages, one record.
Map the whole estate before choosing a single page.
The crawler walks every link in a real browser, lists every page and document, and groups pages that share a template. You approve a scope that covers every distinct template and every journey that matters, and pay for what is distinct, not for repetition.
- Single-page apps followed through their route changes
- PDFs and Office files listed for the document audit
- Critical journeys always in scope, whatever the template count
90,000+ pages inventoried across 1,100+ sites
Agents do what your customers do.
Agents tab through forms, type, open every menu and dialog, submit to bring up errors, and follow signed-in journeys inside your own browser session, so no credentials are handed over. Log-in, OTP and payment controls are never pressed unless you allow it.
- Keyboard order and focus recorded step by step
- Dialogs judged from a timeline of keys, focus and page changes
- Every submit answered locally, so nothing reaches your servers
Watch an agent work in the scene below.
One button, read five ways.
A scan reads the code. The platform stores each screen state with its pixels, its code, the accessibility tree that assistive technology reads, measured contrast and focus, and on apps the exact words the screen reader speaks. A barrier that only shows when the layers disagree is still found.
- Every screen stored, so findings can be re-checked without your live site
- Focus visibility measured from pixels, not assumed
- Android and iOS read element by element, with real screen-reader speech
All 55 criteria, on every screen, with a reason.
Each screen gets a position on every WCAG 2.2 Level A and AA criterion: pass, fail, not applicable, or needs a human. The platform can say "not evaluated"; it never guesses. Specialists judge what machines cannot, and severity is graded on the barrier a person meets, not on the rule that fired.
- One fixed rule-book per criterion, built from measured verdicts
- Every decision carries the sentence that justifies it
| Screen | 1.1.1 | 1.3.1 | 1.4.3 | 1.4.11 | 2.1.1 | 2.4.7 | 3.3.1 | 4.1.2 | 4.1.3 |
|---|---|---|---|---|---|---|---|---|---|
| Log-in | P | P | P | F | P | P | P | P | NA |
| KYC step 2 | P | F | P | P | P | R | P | P | F |
| Order entry | P | P | F | P | F | F | NA | F | P |
| Portfolio | F | P | P | P | P | P | NA | P | P |
Evidence your developers can act on, and an auditor can check.
Every finding is one screen, one element and one criterion, with the screenshot, the code, what a person meets, and the fix. A second reviewer tries to disprove it before it ships; what cannot be reproduced on the stored screen is never reported as confirmed.
- Screenshots rendered from stored evidence, with personal data blurred
- Exported to your tracker as a fixed-column register
- Element
- #mobile-error
- Who it stops
- Blind and low-vision customers completing KYC
- Fix
- Announce the message: role="alert" on the error container
<div id="mobile-error">
Enter a valid…
</div>
Signed reports, and "fixed" that means fixed.
Reports come in the format your regulator or buyer sets, with a position on every criterion, signed and verified before they leave us. On a re-audit every past finding is matched to the current screen and given one status; a finding a scanner stopped raising is re-measured before it is called solved. Monitoring sends each new release back through capture.
- Re-audits of nearly 10,000 findings at a time
- The next annual audit holds no surprises
461 past findings
55 criteria
Watch an agent work
An agent testing a real customer journey.
In this scene an agent tabs through an account-opening form, types, and submits it. The error that appears is shown on screen but never announced to a screen reader, so the agent records a failure of WCAG 4.1.3. On a phone, the screen-reader focus lands on the Buy button and hears only "Button", a failure of 4.1.2. The quantity field reads correctly and passes. Nothing is sent to your servers.
Scene 2 · A trading portal, by keyboard and screen reader
Here the agent uses only the keyboard on a trading screen. Numbered badges show the tab order and the speech panel shows what the screen reader says. The price chart is announced only as "graphic", a failure of WCAG 1.1.1. The quantity field is announced with its name, role and value, and passes. Pressing Buy opens a confirmation dialog, but nothing is announced and the next Tab lands on the Sell button behind it, a failure of 2.4.3.
Your reports, in one place.
Every audit, re-audit and report we deliver is published to the EnableUser client portal. Your team opens it from a private link, with nothing to install.
Open the client portalModules
What runs on the platform.
Web capture
Each screen state stored with its structure, pixels, reading order, tab order, contrast and motion. An industry rule engine and our own pixel, keyboard and focus checks run on the same capture, with a second engine available to cross-check.
Signed-in journeys
Log-in, two-factor, e-KYC, account opening, order entry and dashboards, captured inside your own browser session. Forms are submitted in a sandbox, so nothing reaches your servers.
Android and iOS
The full accessibility tree, focus order and what the screen reader announces, element by element. Text colours read from the running app where the framework allows.
Documents
PDFs, Word, Excel and PowerPoint audited before and after, with the source files corrected so the next export stays accessible. Documents and pricing
Quality scanning
The same crawl checks spelling, broken links, visual layout problems such as overlapping or cut-off text, and page performance, reported alongside accessibility.
Security readiness
How apps handle data, sessions and devices, checked on the same captured record, so weak points are closed before your formal cyber audit.
Quality controls
Checks that are built to fail.
Adversarial review
A second reviewer tries to disprove every finding before it ships. What does not survive is dropped or marked for human review.
Measured precision
We measure how often each kind of finding is right. A class below 90% precision never reaches a client without a second reading.
Controls that can fail
Every gate is tested on a sample built to break it. A check that cannot fail proves nothing, so we do not use one.
See the platform on your own journeys.
Tell us which website, app or portal matters most. We will show you what the agents find on it.
Talk to us →