The Accessibility Engine for automated testing of HTML-based user interfaces. Drop the aXe on your accessibility defects!
Thanks for helping us to make axe-core the most widely used HTML accessibility testing library in the galaxy! Share your axe-core story using #axecoresquad or by submitting it here to get a sweet T-shirt.
We believe that automated testing has an important role to play in achieving digital equality and that in order to do that, it must achieve mainstream adoption by professional web developers. That means that the tests must inspire trust, must be fast, must work everywhere and must be available everywhere.
- Automated accessibility testing rules must have a zero false-positive rate
- Automated accessibility testing rules must be lightweight and fast
- Automated accessibility testing rules must work in all modern browsers
- Automated accessibility testing rules must, themselves, be tested automatically
The sharp edge of the aXe
aXe is the third generation of accessibility rules for HTML-based user interfaces that differentiates itself from other approaches and rules repositories in the following ways:
- It works on all modern browsers,
- It supports in-memory fixtures, static fixtures, integration tests and iframes of infinite depth
- It has zero false positives (bugs notwithstanding)
- It is open source
- It is actively supported by a major accessibility vendor
- It is designed to work with whatever tools, frameworks, libraries and environments you have today
- It is designed to be integrated into your existing functional/acceptance automated tests
- It automatically determines which rules to run based on the evaluation context
- It is highly configurable
First download the package:
npm install axe-core --save-dev
Now insert calls at each point in your tests where a new piece of UI becomes visible or exposed:
The aXe API fully supports the following browsers:
- Microsoft Edge v40 and above
- Google Chrome v42 and above
- Mozilla Firefox v38 and above
- Apple Safari v7 and above
- Internet Explorer v9, 10, 11
Support means that we will fix bugs and attempt to test each browser regularly. Only Firefox and Chrome are currently tested on every pull request.
There is limited support for JSDOM. We will attempt to make all rules compatible with JSDOM but where this is not possible, we recommend turning those rules off. Currently the
color-contrast rule is known not to work with JSDOM.
The Accessibility Rules
The complete list of rules run by axe-core can be found in doc/rule-descriptions.md.
Contents of the API Package
The aXe API package consists of:
axe.min.js- a minified version of the above file
Axe can be built using your local language. To do so, a localization file must be added to the
./locales directory. This file must have be named in the following manner:
<langcode>.json. To build aXe using this locale, instead of the default, run aXe with the
--lang flag, like so:
grunt build --lang=nl
This will create a new build for aXe, called
axe.<lang>.min.js. If you want to build localized versions, simply pass in
To create a new translation for aXe, start by running
grunt translate --lang=<langcode>. This will create a json file fin the
./locales directory, with the default English text in it for you to translate. We welcome any localization for axe-core. For details on how to contribute, see the Contributing section below.
To update existing translation file, re-run
grunt translate --lang=<langcode>. This will add new messages used in English and remove messages which were not used in English.
Additionally, locale can be applied at runtime by passing a
locale object to
axe.configure(). The locale object must be of the same shape as existing locales in the
./locales directory. For example:
Supported ARIA Roles and Attributes.
Refer aXe ARIA support for a complete list of ARIA supported roles and attributes by axe.
Read the Proposing Axe-core Rules guide
Read the documentation on the architecture
Read the documentation on contributing
Projects using axe-core
Thanks to Dulin Marat for his css-selector-parser implementation which is included for shadow DOM support.
Thanks to the Slick Parser implementers for their contribution, we have used some of their algorithms in our shadow DOM support code.