TestAlly for Developers

Your component’s personal manual testing guide

AXE tells you what’s wrong in automated tests (and catches about 35% of the issues), ChatGPT guesses how to fix it.

TestAlly does both, with vetted WCAG-backed sources and minimal code diffs. The difference with all other tools out there is that we offer a per-component manual testing walkthrough that developers can trust.

TestAlly aims to be a manual testing assistant that bridges the gap where automated accessibility tools fall short. Accessibility is still a topic of great worry and overwhelm for developers, and this tool makes their lives much easier.

Using AI to interpret a component’s intent and code, TestAlly will provide developers with immediate, step-by-step manual testing instructions or a definitive “all-clear” if no testing is required. No more waiting on external accessibility teams to ship accessible code; TestAlly will provide guidance on each test.

The tool aims to create a self-supporting workflow that teaches the “why” behind accessibility with context and turns manual verification from a bottleneck into a real-time confidence check.

TestAlly distinguishes itself from traditional tools like Accessibility Insights, tota11y, Storybook a11y, and Pa11y by transforming the manual testing process from a generic requirement into an AI-guided, component-specific workflow.

Target Audience

We’re looking for contributors who have particular skills in

AI prompting
Backend development (API integrations and code analysts)
Frontend development (Next.js, React, general UI development)
UX/UI – specifically around design for the interface, but also accessibility
Technical writing for documentation
Quality Assurance – for validation and testing
Accessibility expertise (WCAG/ARIA Standards, component patterns, etc)

Hackathon Goals

Build a working MVP
This will accept HTML, JS, and CSS as an input and generate manual testing instructions based on the component type
Implement custom detection rules
These are the minimum two cases we want to get working.
- Link-as-button detector: <a> without href but with onClick.
- Focus ring removal detector: flags CSS removing outline without visible replacement.
- Stretch Goal: Context detection simply from the input component and CI (GitHub action) implementation to add a comment on PR with test scenario and explanations.

Create comprehensive documentation
Explain the tool in detail, its usage, and possible extensions/development
Create a project website
Opportunity to showcase TestAlly, thought process, background to the project, and current status, along with a community-based option (GitHub?) to support future development

Results

TestAlly tackles the gap that catches developers off guard: automated accessibility tools cover roughly 30 to 40 percent of issues. The rest, being keyboard navigation, screen reader behaviour, and voice control compatibility, requires manual testing that most developers were never taught.

Over the hackathon weekend, our team built a working AI-powered tool that closes that gap. Paste in your HTML, CSS, and JavaScript, describe what the component is supposed to do in one sentence, and TestAlly returns either a structured set of manual testing instructions for keyboard, screen reader, and voice control, or a clear all-clear. Every recommendation cites its source. The tool does not guess.

The real challenge was integration: getting all the moving parts to talk to each other within 48 hours. AI could support parts of the build, but connecting everything into a working whole was on us. We are proud of what the team delivered: a working MVP.

What’s next: a GitHub Action that posts test scenarios automatically on pull requests, and context detection so developers no longer need to describe the component manually. Contributions are welcome — code, testing, documentation, translation.