Ann Arbor was originally an internal project at FormulaFolios and was a collaborative project by multiple contributors. It was part of an ongoing project to better spot, fix, and improve accessibility issues across our sites and clients that we were serving. We made it an open source project as part of our Doing Good initiative as there weren’t as many other low cost options for spotting and fixing accessibility violations besides WAVE API.

It uses Dequeue Axe Core API to crawl and audit sites for accessibility violations.

The original repo can be found on https://github.com/formulafolios/ann-arbor but the old team is now defunct and I am the sole maintainer of the fork on https://github.com/bennpham/ann-arbor.

The readme should do a good job explaining on the requirements and functionality of the tool, but ultimately it is a command-line tool that uses Python to crawl and audit a single page or entire site when given a URL or domain. Not all accessibility violations will be found by the tool, but this tool will show you which page have the most errors to allow auditors to priotize which pages to improve on first.

You’ll need to install Python and Chromedriver then you can audit the page via the following commands:

Organize by template (URL path):

Organize by page:

Single page auditing:

Specific report type (color contrast violations only):

Specific report type (all violations EXCEPT color contrast):

You’ll get a handful of json files and a csv file of all of the violations that you have in your audits/httpbin.org/ folder or whatever domain you’re auditing on.

I basically just kept this package up to date as I still use it myself here and there to help out some friends audit their sites for any accessibility violations on top of helping them with SEO related tasks.

Originally the latest version was on Python 3.6.0. The final security fix was Python 3.6.15 and Python 3.6 reached EOL on September 4, 2021.

The latest pull request I made was to update to Python 3.12 which should last until October 2028. The original test suite was nose which was long abandoned and the standard for Python testing is pytest.

Lastly, the original PR requires the internal teams of FormulaFolios to manually run their tests and linters in order to make sure that the code is working and coding best practices are adhere to. There isn’t a strict safeguard in case any test or linter rules fail. Back then, we did not have GitHub CI but there was TravisCI or CircleCI that you can use to handle merging rules. For a smaller project we didn’t utilize those. I added GitHub Actions CI if anyone else wants to contribute to the project. That’ll make it easier to follow a coding guideline to collaborate with multiple people when our linter rules and test pass rules are enforced.

Also a lot of thanks goes to Alicia Evans who was the site accessibility engineer during my time at FormulaFolios who made this project possible.

As it is, Project Ann Arbor is nice enough as a dev internal use tool. For the average non-technical user, they probably wouldn’t know where to start using it as it’s a command line based tool with a few installation steps.

A nice idea could probably be a web interface that a user could go on, input their website, and get a quick list of what they need to improve on while Project Ann Arbor can generate all of the csv file and audits in the backgroudn while the web based front end can parse that and display it to an average user in a friendlier way to let them know what they’d need to audit similarly to typing in your email address on HaveIBeenPwned to check for data breaches.