Would You Trust an AI to Fix Accessibility in Your Codebase?

Accessibility (a11y) is one of those things every dev knows matters—until the deadline hits.

We add alt text later. We tweak color contrast later. We fix ARIA warnings later. And “later” either never comes, or it takes ages of torture to fix. 🤕

I’ve been experimenting with a dev tool idea called AccessFix—a GitHub-integrated service that automatically detects accessibility issues in your pull requests and creates a PR with the fixes already applied.

It’s like having an a11y engineer review every commit—without extra process overhead.

Before I build this into a real SaaS, I want to ask the community:

• Would you trust AI-generated accessibility fixes?
• What’s the most annoying part of making your app accessible today?
• Would your team use this if it came as a GitHub App or CI plugin?

Curious to hear from anyone who’s had to juggle performance, deadlines, and accessibility at once, or just had to suffer through ages of checking mundane a11y errors.

Let’s make the web less broken. 🌐