How AuditAble Works
Automated accessibility scanning, practical remediation guidance, and honest coverage boundaries.
How Our Scanning Works
WCAG 2.1 AA has 50 success criteria. No single tool tests all of them. Here is how we stack up.
Layer 1: Automated Browser Scan (28 of 50 criteria)
All PlansWe load your live pages in a real browser with Playwright and run axe-core against the rendered DOM, then operate the page like a resident would: tabbing through focusable elements to catch keyboard traps, unreachable controls, and invisible focus, resizing to a 320 pixel viewport for reflow, and reading the browser accessibility tree the way a screen reader does. Every scan records exactly which criteria its rules evaluated, 28 of the 50 WCAG 2.1 AA success criteria today.
Static axe-core rules cover 14 criteria, the same engine most platforms stop at. Our keyboard and accessibility-tree checks cover 14 more, and we only count a criterion as tested when a rule for it actually ran.
Layer 2: AI-Assisted Evidence + Screening
In ValidationOur scanner captures page metadata for AI-assisted review of judgment-heavy items such as alt text quality, link text quality, heading quality, form-label quality, consistent navigation, consistent identification, and reading order. It also screens collected media and interaction evidence, video without captions, autoplaying audio, flashing animation, and flags likely problems for review.
When Layer 2 runs, the tested count rises to 33 of 50 criteria, and screening flags likely issues on more without marking them tested. Criteria we did not test are reported as Not Evaluated, never as passed.
Layers 3 & 4: Multimedia + Advanced AI
RoadmapThese are deeper roadmap layers, not current plan-gated coverage. The scanner now captures media and interaction evidence; next we will add automated caption-quality checks with speech-to-text, OCR, and timestamp comparison, plus images-of-text identification and deeper complex-ARIA validation. These reduce manual review burden without replacing human review.
Remaining Criteria: Requires Human + Screen Reader
Manual ReviewSome criteria require a person using NVDA, JAWS, or VoiceOver: whether dynamic content changes are announced coherently, whether form error recovery is navigable by assistive technology users, and certain timing-related judgments. We mark those criteria honestly instead of counting untested items as passes.
How We Compare
| Vendor | Automated Coverage | Key Difference |
|---|---|---|
| Most tools | ~30-35% | Static scanning only, often using the same core engines |
| Overlay vendors | ~35% + overlay | Adds JavaScript overlay instead of fixing source issues |
| Manual audit firms | ~35% + manual | Useful but expensive, point-in-time, and hard to repeat often |
| AuditAble all plans | 28 of 50 criteria, with an evidence ledger | Rendered-page axe-core scan plus a per-criterion ledger that names what was tested, what was flagged for review, and what still needs a person |
| AuditAble with Layer 2 (in validation) | 33 of 50 criteria, plus AI review flags | AI-assisted evaluation of judgment-heavy criteria; multimedia and manual partner workflows will be labeled when live |
We do not report criteria we could not evaluate as passed. Coverage is the same across plans; plans differ by capacity, reporting, team seats, and workflow features.
What You Can Test Yourself
Your scan report includes manual checks we cannot automate, with step-by-step instructions.
Screen Reader Navigation
Open NVDA, JAWS, or VoiceOver. Navigate priority pages using only the keyboard and listen for heading order, field labels, and dynamic updates.
Video & Audio Content
Confirm videos have captions, captions are accurate, and visual-only content has an audio description or equivalent text.
Keyboard-Only Navigation
Put the mouse away. Use Tab, Enter, Escape, and arrow keys to complete important tasks.
Zoom & Reflow
Zoom to 200% and confirm text reflows without horizontal scrolling, overlap, or hidden controls.
Why Now
- DOJ Title II deadline: April 26, 2027 for entities serving 50K+ population; April 26, 2028 for all others.
- Enforcement is real: DOJ actions and private complaints make accessibility risk practical, not theoretical.
- Proactive is cheaper: Continuous monitoring costs a fraction of a one-time manual audit or lawsuit response.
What We Don't Do
Honesty matters, especially with government buyers. Here is what AuditAble is not:
- We do not provide legal advice or certify compliance. Those are legal determinations.
- We do not touch your servers or implement code fixes. We provide guidance; your team deploys.
- We do not claim to test 100% of WCAG criteria. Automated tools cannot.
- We do not do real-time monitoring. We run scheduled periodic scans.
See Where You Stand
Run a free scan on your website. It takes a few minutes, no account needed.