Web accessibility auditing is time-consuming and inconsistent when done by hand. A single website can contain hundreds or thousands of pages, each requiring systematic inspection against dozens of technical criteria. Organisations face mounting legal exposure from non-compliance under ADA, WCAG 2.1, Section 508, and the European EN 301 549 standard — yet most lack the tooling to monitor their sites continuously, track remediation progress over time, or produce the audit-ready reports that legal and compliance teams actually need.
Accessibility SaaS
Ablelytics
Automated web accessibility testing at scale.
A cloud-based SaaS platform that automates accessibility auditing across entire websites, maps violations to WCAG 2.1, ADA, Section 508, and EN 301 549, and produces professional PDF reports — turning what would otherwise be a slow manual process into a continuous, scalable workflow.
4
Compliance standards
1,000s
Pages per scan
4
Pricing tiers
Cloud SaaS
Deployment
The Problem
Manual accessibility auditing doesn't scale.
Key challenges
- A single manual audit can take weeks for a large site and is outdated the moment development resumes
- Legal risk is increasing: ADA litigation has risen sharply year on year across the US and Europe
- Accessibility regressions are introduced silently by routine code changes and content updates
- Compliance teams need evidence trails and structured reports — not raw code findings
What We Built
Continuous automated auditing with report-ready output.
Ablelytics is a three-part platform: a Next.js dashboard where users manage projects, configure scans, and access reports; a standalone Node.js scanning worker that launches headless Chromium via Puppeteer and runs axe-core against every page; and a Sanity-backed public website. The dashboard supports on-demand and scheduled scans, real-time progress via Server-Sent Events, a searchable report library, team accounts, Stripe billing, and a full REST API for CI/CD integration. The scanning worker processes up to 100 pages concurrently, stores granular violation data in Google BigQuery for historical analysis, and generates professional PDF audit reports covering executive summaries, severity breakdowns, and rule-by-rule analysis.
Capabilities
What the platform does
Project and page management
Add websites, organise pages into custom groups, and define include/exclude rules using regex or wildcard patterns. Page sets let teams focus scans on specific sections of a site.
Scan orchestration
Trigger full-site scans on demand or schedule recurring scans — daily, weekly, or monthly. The worker processes pages concurrently, respects robots.txt, and enforces configurable depth and page-count limits.
Real-time progress
Scan status streams to the browser via Server-Sent Events. No polling, no page refresh — live updates throughout the scan pipeline from page collection through to report generation.
Professional PDF reports
Each completed scan generates a PDF audit report covering executive summary, severity distribution, and per-rule analysis with affected elements, CSS selectors, and screenshots. Suitable for direct delivery to clients or compliance teams.
API access
Manage bearer tokens, configure webhooks, and integrate scan triggering into CI/CD pipelines. API call logs with cursor-based pagination provide a full audit trail of programmatic usage.
Team collaboration
Organisation accounts with role-based access. Multiple team members can share a project, review scan results, and manage subscriptions without separate accounts.
Technical Architecture
Key decisions and why
01
Migrated from Firestore to PostgreSQL in nine production-safe phases
The platform was originally built on Firebase/Firestore. As data complexity grew, the lack of relational integrity, unpredictable query costs, and limited schema control became constraints. We migrated the entire data layer to PostgreSQL with Prisma ORM across nine incremental phases — each targeting a specific domain (auth, projects, pages, subscriptions, reports, worker API) — with no downtime and no data loss.
02
Server Actions and Server Components throughout
We use Next.js 15/16 patterns throughout: server components fetch data at request time, server actions handle mutations, and client components are reserved for interactive UI. This minimises client-side JavaScript, eliminates redundant API routes, and keeps sensitive database logic on the server.
03
SSE over WebSockets for real-time scan updates
Scan progress and notifications are streamed to the browser via Server-Sent Events rather than WebSockets. SSE is unidirectional — which is all we need — and avoids the operational complexity of a stateful WebSocket server. It works through HTTP/2 multiplexing and is trivially compatible with serverless deployments.
04
Cursor-based pagination across all API endpoints
API logs and paginated data use cursor pagination rather than offset pagination. Cursor pagination is stable under concurrent writes and performs consistently even across large datasets — offset pagination degrades as the dataset grows and produces inconsistent results when rows are inserted or deleted.
Technology
Stack
Frontend
Next.js 16.1React 19TypeScript 5Tailwind CSSStyled-components
Backend
Next.js API RoutesServer ActionsNode.jsPrisma ORM 5
Database & Storage
PostgreSQL 16Google BigQueryGoogle Cloud Storage
Scanning Worker
Node.jsPuppeteer 24axe-core 4pdfmake
Auth & Payments
Auth.js (next-auth v5)Google OAuthStripe
CMS & Deployment
SanityFirebase HostingVercel
Results
Outcomes
- 01
Automated scan pipeline capable of processing thousands of pages per run with concurrent page execution
- 02
Professional audit-grade PDF reports suitable for direct stakeholder delivery
- 03
Continuous monitoring with scheduled scans and email notification workflows
- 04
Full API surface enabling integration into development and CI/CD pipelines
- 05
Secure multi-tenant architecture with organisation-level data isolation and role-based access control
- 06
Four-tier subscription model with Stripe billing, annual/monthly plans, and a 14-day free trial
Work with us
Building something similar?
We bring the same depth of engineering to client work as we do to our own products.