Why Clamper
96.3% of top-million home pages have detectable WCAG errors (WebAIM Million 2025). EAA enforcement is active in the EU since 28 June 2025. Yet most CI/CD pipelines have no accessibility gate. The reason is structural — not a deficiency of any underlying scanner.
The CI/CD a11y deadlock
Existing tools answer "is any violation present?" — not "did this PR introduce any new violation?" Teams with pre-existing technical debt (which is everyone with a real codebase) face a binary choice:
- Block every commit — every PR fails on debt that pre-dates the change. Devs disable the gate within days.
- Disable the gate entirely — no regression detection at all; new violations slip into production.
Both outcomes leave the organisation exposed. The deadlock is structural.
Existing CI a11y tools
Source: GitHub stars + first-release dates per each project's repository, sampled 2026-04-29.
| Tool | First release | Baseline diff? | Policy DSL? | Exemption ledger? | Multi-domain? |
|---|---|---|---|---|---|
| Pa11y CI | Jun 2016 | No | Flat config | No | a11y only |
| Lighthouse CI | 2019 | No | Static threshold | No | a11y + perf |
| cypress-audit | Feb 2020 | No | None | No | a11y + perf |
| Equa11y | Aug 2020 | No | None | No | a11y only |
| Deque axe Linter (closed beta) | 2024 | No | Org-flat | Manual file-list | a11y only |
| Clamper | 2026 | Yes | YAML cascade | Yes (with expiry + DOM invalidate) | a11y + privacy + security |
Five validated pains
Per medium customer (50 developers, ~30 repos). Source citations below.
| # | Validated pain | Annual cost without Clamper | Savings with Clamper |
|---|---|---|---|
| 1 | A11y regressions reach production (96.3% sites have errors per WebAIM Million 2025) | ~$100K incidents + hotfix MR cycles | ~$75K saved (block at the gate) |
| 2 | Per-repo bespoke pipeline glue maintenance | ~$60K (2 platform-eng hours/wk × 50 wk) | ~$45K saved (single org-level policy) |
| 3 | EAA / EU AI Act audit reconstruction | ~$20K manual log assembly per quarter | ~$18K saved (signed audit chain per build) |
| 4 | Disabled gates after dev revolt (impractical strict baseline) | ~$50K regressions × N + lawsuit risk | ~$40K saved (gradient ramp + delta tolerance) |
| 5 | ADA / EAA litigation amortised (4,605 lawsuits in 2024) | ~$20K/yr amortised across customers | Liability reduced by recorded gate enforcement |
External evidence
- WebAIM Million 2025: 96.3% of the top one million home pages have detectable WCAG errors.
- Seyfarth Shaw 2024: 4,605 ADA Title III accessibility lawsuits filed in 2024; 69% targeted e-commerce.
- EAA — Directive (EU) 2019/882: active since 28 June 2025 across all 27 member states. Per-state penalty schedules (DE €100K, FR €250K, SE €200K).
- EU AI Act Article 50: AI-generated content marking mandatory August 2026; non-compliance penalty up to €35M or 7% global turnover.
- ADA Title II web rules: phase in across April 2026 (large public entities) and April 2027 (small).
- EN 301 549 v3.2.1: European harmonised standard underlying EAA conformance; codifies WCAG 2.1 AA plus non-web ICT requirements.
Why now
Continuous attestation — auditable, signed evidence that every commit was checked — is now an unmet need. Quarterly Siteimprove PDFs are stale by the time they ship. Disabled Pa11y CI gates leave a blind spot. Clamper's hash-chained signed audit record is the regulator-acceptable artefact, generated as a side-effect of every gate decision.