AI Learning

How Memory learns from your usage patterns

AI Learning

Memory automatically learns from how you use Metrifyr tools, creating personalized context that improves your experience over time.

What It Learns

Learning from Mistakes

When you make API errors, Memory learns to help you avoid them in the future:

• Wrong property format - Reminds you of the correct format next time
• Missing required parameters - Suggests your common defaults
• Permission errors - Notes which accounts have proper access

This means the AI gets smarter about preventing errors you've encountered before.

Viewing Learned Items

1. Go to Dashboard > Memory
2. Select your API key from the dropdown
3. Look for items marked "Learned" in the source column

Managing Learned Items

You have full control over learned items:

• Keep useful items - They improve over time with continued use
• Delete incorrect items - Click the trash icon to remove
• Deactivate temporarily - Toggle off without permanently deleting

Priority System

Memory items have priority (1-100). Higher priority items are used first when the AI needs context.

Automatic Priority Adjustment

• Frequently used items get priority boost
• Unused items gradually decrease in priority
• Learned items with very low priority are automatically deactivated

This ensures your most relevant context is always prioritized, while unused items fade away naturally.

Triggering Learning

Learning happens automatically in the background. You can also trigger it manually:

1. Go to Dashboard > Memory
2. Select your API key
3. Click "Analyze Now"

Manual analysis is useful when you want immediate learning after significant usage.

Best Practices

• Use tools consistently - The more you use tools with the same parameters, the better learning works
• Review learned items - Periodically check what was learned and remove incorrect items
• Keep memory enabled - Learning only works when memory is active for your API key