How to use accessCode chat commands in GitHub

Note: accessCode is currently in beta and only available to eligible customers.

The accessCode AI Reviewer is more than just a static analysis tool. It is an interactive, conversational accessibility expert that works directly inside GitHub.

You can communicate with the bot through GitHub comments using native slash commands. This helps you understand issues, explore fixes, and control how the AI Reviewer behaves, all without leaving GitHub.

How chat commands work

Interact with the bot by typing commands in pull request comments.

Each command starts with:

/accessCode

Where you post the command matters. The bot uses different context depending on the comment type.

Comment types

Top-Level comments (for PR-wide actions)

Post these in the Conversation tab of the pull request (PR). Use top-level comments when you want to control the AI reviewer or run actions across the entire PR. For example, requesting a new review, or pausing the reviewer.

Thread replies (for code-specific help)

Reply directly to a comment the bot created in the Files changed tab. Use thread replies when you want the bot to analyze or modify a specific issue in your code. The bot uses the thread context to answer.

Global commands (Top-level)

Global commands are posted as top-level comments on the PR Conversation timeline.

To use them, type /accessCode followed by the command:

Command	What is does
`/accessCode help`	Shows all available commands and how to use them.
`/accessCode review`	Immediately runs a new scan of the entire PR diff. Useful if you’ve updated `.accesscode.yml` configuration and want to apply changes without pushing a new commit.
`/accessCode pause`	Pauses the AI Reviewer from scanning new changes in this PR. The bot will acknowledge the pause and stop listening to further commits or comments, saving resources if you are pushing rapid WIP updates.
`/accessCode resume`	Resumes the AI Reviewer and immediately runs a new scan of the PR.

Contextual commands (thread replies)

These commands are conversational and use the flagged line of code to provide more relevant answers. Contextual commands must be posted as Thread Replies to an existing accessCode finding.

Command	Action
`/accessCode explain`	Explains why the issue was flagged and how it affects accessibility.
`/accessCode alternative`	Suggests a different way to fix the issue. Useful if the bot’s initial suggested fix breaks your UI, violates your design system, or simply isn’t what you want.

What to expect after sending a command

After you send a command, the bot reacts with an 👀 emoji. This means your command was received and the bot is processing your request. Some responses may take a few seconds, especially for more complex code analysis. Once ready, the bot replies with the result.

accessiBe Help Center