WCAG Testing Tools: 11 Compared by Coverage, Cost, and Workflow Fit [2026]

Last updated: November 12, 2026

WCAG testing tools split into three categories — free engines and extensions, CI/automation runners, and commercial platforms — and every one of them, regardless of price, tops out around the same detection ceiling. This guide compares 11 tools by category, not rank: what each tests, where coverage stops, its cost model, and the manual layer none of them replace.

Key numbers: Automated detection covers roughly 60–70% of WCAG issue volume across every tool here, free or commercial, with the remaining 30–40% requiring expert human review (TestParty remediation data across 100+ brands, as of September 2026). WCAG 2.2 defines 55 success criteria at Levels A and AA — the fixed target every tool below is measured against (W3C). The open-source axe-core engine behind most tools on this list maps 74 rules to WCAG plus 27 best-practice rules (GitHub). In 2026, 95.9% of the top one million home pages carried a detectable WCAG failure, averaging 56.1 errors per page (WebAIM Million, 2026). axe DevTools Pro is reported at $1,200–$2,500 per seat annually (Vendr, as of September 2026).

TestParty competes in this market, including with the free scanner and Bouncer CI gate named below. Tools are grouped by category, not ranked — this comparison uses public information and cited sources; evaluate every option against your own requirements.

How should you compare WCAG testing tools?

Compare WCAG testing tools on five criteria — engine and rule quality, how much of the 55 Level A/AA criteria a tool actually reaches, workflow fit, whether the output is a finding or a fix, and cost — never on brand recognition alone.

• Engine and rule quality. Most tools here run axe-core or a comparably mature open-source engine — you're rarely choosing a different rules library, just a different wrapper around one.
• Coverage of the 55 A/AA criteria. No tool automates more than a minority of them. Ask which criteria a tool marks "pass/fail" versus "needs review."
• Workflow fit. A one-page extension, a CI gate, and a scheduled crawl answer different questions.
• Output actionability. Ten of the eleven tools below end in a report someone has to act on. One turns findings into a pull request.
• Cost. Free-forever, per-seat commercial, and quote-based enterprise — price tracks workflow scale, not detection depth.

Which free engines and extensions should you know?

Six free tools carry most of the market's detection work, and five of the six share axe-core or a comparably built open-source engine underneath a different interface.

axe-core / axe browser extension. axe-core is the MIT-licensed, open-source engine most of this list runs on, maintained by Deque Labs and built for near-zero false positives; the free axe DevTools extension is its manual, one-page interface. Strength: the engine standard — Lighthouse, Microsoft's tooling, and TestParty's own stack all embed it. Blind spot: detects only. Cost: free. Best for developers wiring tests directly into a codebase.

WAVE. Built by the nonprofit WebAIM, WAVE overlays findings directly on the page — icons on an unlabeled image, brackets around an empty form field. Strength: the fastest way to see where a violation lives. Blind spot: one page per run, no CI hook. Cost: free.

Lighthouse. Built into Chrome DevTools, Lighthouse scores accessibility 0–100 from a subset of axe-core's rules and scripts into CI. Strength: a fast regression signal. Blind spot: a 100 means the automatable checks passed, not WCAG conformance. Cost: free.

Microsoft Accessibility Insights. Accessibility Insights is axe-core-based like the three above, but its Assessment mode adds a structured, guided walkthrough of criteria a person has to check by hand. Strength: bridges automated findings and a real manual pass. Blind spot: a point-in-time desktop tool, not a CI gate. Cost: free.

ANDI. Built by the Social Security Administration for the federal Trusted Tester process, ANDI is a free, open-source bookmarklet that reveals what a screen reader would announce for an element — its computed accessible name, role, and tab order — overlaid on the page. Strength: surfaces name computation and focus order more directly than any tool above. Blind spot: manual, one page at a time, no dashboard. Cost: free (Section508.gov lists it in the federal toolkit).

IBM Equal Access Accessibility Checker. IBM's Apache-2.0 open-source checker is a genuine axe-core alternative — its own engine mapped to WCAG 2.0–2.2, shipped as browser extensions plus a Node package wrapping Selenium, Puppeteer, Playwright, Cypress, and Vitest. It holds this slot over the W3C's Nu Html Checker, which validates HTML syntax rather than WCAG criteria — useful for markup errors, not accessibility rules. Strength: CI coverage rivaling axe-core, from an independent engine. Blind spot: smaller ecosystem, still detection-only. Cost: free, open source.

Which CI and automation tools fit a dev pipeline?

Two tools turn detection into a pipeline step rather than a one-off scan, and they split on what happens after a violation is found: one produces a finding, one produces a pull request.

Pa11y / pa11y-ci. Pa11y wraps a rules engine into a CLI and CI-ready reporter at zero cost, and both Pa11y and pa11y-ci remain active open-source projects as of 2026. Strength: lightweight and scriptable, no per-seat license. Blind spot: community-maintained, not commercially backed — support is whatever the project's contributors provide. Cost: free.

TestParty Bouncer + free scanner. Disclosure: TestParty builds this product, so weigh it accordingly. Bouncer gates CI/CD on new violations the way a `@axe-core/playwright` check would; the free Shopify scanner is the entry point for template-level coverage. What it adds over every other tool here: detection wired to source-code remediation rather than a dashboard — 35 million+ issues remediated as of September 2026, 14-day initial remediation, customers typically spending 15–30 minutes a month reviewing the resulting pull requests. Cost: free scanner app; remediation scoped per store. See accessibility in CI/CD for the fuller pipeline.

Which commercial platforms should you evaluate?

Three platform-class options add scale, workflow, or governance on top of the same detection ceiling as the free tools above — none of them changes what a rules engine can prove.

axe DevTools Pro (Deque). Pro is a self-serve, per-seat subscription layered on axe-core, adding AI-enhanced guided tests, deduplication, and Jira integration. Deque publishes no list price; procurement data reports negotiated seats at $1,200–$2,500 per year (Vendr, as of September 2026). Strength: a better scanning UI for teams with capacity to act on a dashboard. Blind spot: the same 60–70% ceiling as the free extension underneath it. See Deque axe alternatives for dev teams for what replaces this tier.

ARC Toolkit / TPGi platform-class tools. TPGi, a Vispero company in the same family as the JAWS screen reader, offers ARC Toolkit free for single-page testing, plus a quote-based ARC Platform for site-wide monitoring and audit-workflow management, alongside manual audit and VPAT services. Strength: pairs tooling with certified expert audits under one vendor. Blind spot: the platform still detects at the same rate as any other tool here, and an audit is a snapshot that starts aging on delivery. Cost: free extension; platform and audits quote-based.

Siteimprove-class monitoring suite. This category bundles accessibility scanning into a broader digital-governance platform alongside content quality, SEO, and analytics, sold on scheduled crawls and dashboards. Reported contracts commonly range roughly $15,000–$150,000+ a year by page volume and module count (Vendr, as of August 2026). Strength: one dashboard covering many properties. Blind spot: monitoring is the map, not the repair — a tracked finding still needs an engineer to fix the code.

How do all 11 tools compare at a glance?

The table below sorts every tool by category, engine, and — the line that matters most — whether its output is something to read or something to merge.

What can't any tool on this list do?

None of these 11 tools closes the gap between automated detection and full WCAG conformance — automation, regardless of price or vendor, catches roughly 60–70% of issue volume and stops there.

In TestParty's audits across 100+ brands, stores already running three or four tools from this list still miss the same failures — the gap was never about which scanner runs, but what a rules engine can prove. The remaining 30–40% concentrates in failures no rules engine evaluates: whether alt text describes the product or just exists, whether keyboard focus follows a real order, whether a screen reader announces a cart update when it changes. Closing that takes two things no tool above provides — assistive-technology testing with NVDA or VoiceOver, and a recurring manual audit, run monthly, checking what automation only flags for review. See manual vs. automated accessibility testing and the methods, tools, and cadence guide.

Which stack should you run?

Run a combination suited to your team, not a single tool — a solo merchant, a dev team, and an enterprise each need a different mix of the 11 above, and none is a complete answer alone.

• Solo merchant or small site. WAVE or Lighthouse for a free baseline, TestParty's free scanner for template-level coverage on Shopify, and ANDI or a manual screen-reader pass before a major launch. A commercial platform is more workflow than a small team needs yet.
• Dev team. axe-core directly in tests (`jest-axe`, `@axe-core/playwright`) and CI, a Lighthouse budget per template, Pa11y or Bouncer gating pull requests, and Accessibility Insights or IBM's checker for the guided manual checks a linter can't run. See automated accessibility testing: 2026 implementation.
• Enterprise. A governance platform — Siteimprove-class monitoring or TPGi's ARC Platform — for scale across many properties, axe DevTools Pro for engineering teams staffed to act on findings, and a monthly manual audit with VPAT documentation for the roughly 30–40% no platform reaches.

None of these combinations makes any single vendor, TestParty included, the right fit for every team. An enterprise already running Siteimprove or ARC Platform for governance doesn't need to replace it; a solo merchant gets more from a free scanner and a monthly manual pass than from any per-seat contract here. For a checker-level first look rather than this professional comparison, see our guide to website accessibility checker tools.

Frequently Asked Questions

What's the single best WCAG testing tool? There isn't one — the three categories answer different questions. A free engine shows what's wrong on the page; a CI tool shows whether a build regressed; a commercial platform adds scale or, in TestParty's case, a path from finding to merged fix. Pick by category and workflow, not brand.

Is axe DevTools Pro just axe-core with a price tag? Mostly. axe-core stays free and MIT-licensed regardless of what you pay Deque. Pro adds a commercial UI, guided tests, and Jira integration on the same engine — see our Deque axe alternatives guide for what replaces that paid layer.

Do any of these 11 tools guarantee WCAG compliance? No. Every tool here evaluates only the portion of the 55 Level A/AA criteria that reduces to a code-detectable condition. A clean scan means the automatable checks passed — not that focus order, alt-text quality, or checkout are actually fine.

Which tool should a non-developer start with? WAVE, because it overlays findings on the rendered page instead of a list of code selectors. ANDI is the next step up for formal 508-style testing, since it shows exactly what a screen reader would announce.

Do I need a commercial platform if I'm already running the free tools? Only if you need what the free layer skips: crawling hundreds of pages on a schedule, a governance dashboard for multiple teams, or certified audit documentation. For detection alone, the free engines above cover most of what a commercial platform's engine does.

How often should these tools actually run? Automated tools belong on every pull request plus a daily production scan. The manual layer none of them replace — NVDA or VoiceOver checks, a full keyboard pass through checkout — needs a person monthly, not annually.

This article was produced using TestParty's cyborg approach — AI-assisted research and drafting, validated and refined by our accessibility team. The analysis above represents TestParty's editorial opinions based on publicly available data. As a competitor in the accessibility market, we have a point of view — but we've cited our sources so you can verify every claim independently.