Providing feedback

mrge’s AI adapts to your team’s coding preferences and standards through feedback.

If you encounter an AI-generated comment that isn’t accurate or relevant, you can respond directly by tagging it with @mrge-io and providing clarification or corrections.

Feedback given on comments triggered by custom rules is stored specifically for those rules, while feedback given on general AI reviewer comments is stored as general feedback.

​

Use cases

• False positives: If the AI flags a pattern that is actually valid for your codebase.
• Team-specific standards: You might have unique naming conventions or special frameworks that the AI should be aware of.
• Rule adjustments: If a custom rule is too strict, you can specify exceptions directly in the comment thread.

​

Viewing and managing feedback

To view or manage the AI’s stored feedback, use the AI review settings → Memory management. You can edit or remove specific items if they become outdated.

​

Additional reactions

You can also vote on AI comments (e.g., thumbs-up/thumbs-down), to guide the AI’s internal ranking of comment quality.