Memory Records

Memory Records help the AI assistant remember your preferences, coding patterns, and instructions across all projects. Tell it once, and it remembers forever.

What Are Memory Records?

Memory Records are persistent notes that the AI assistant uses to personalize its help based on your preferences and past decisions. They're shared across all your projects, so you don't need to repeat instructions.

Memory Types

Preference

Your coding style and tool choices:

• "I prefer TypeScript over JavaScript"
• "Use Tailwind CSS for styling"
• "Prefer functional components in React"

Instruction

Explicit rules about what the AI should always or never do:

• "Never use any type in TypeScript"
• "Always add error handling to async functions"
• "Use async/await instead of .then() for promises"

Pattern

Recognized coding patterns and architecture you follow:

• "I use feature-folder structure for React apps"
• "Follow atomic design principles for components"
• "Prefer composition over inheritance"

Context

High-level background about your projects or work:

• "Building a SaaS application for team collaboration"
• "Working on educational platform for students"
• "Team of 3 developers, using Agile methodology"

Decision

Architectural decisions and their reasoning (lightweight ADRs):

• "Using REST API instead of GraphQL for simplicity"
• "Chose PostgreSQL for relational data requirements"
• "Avoiding Redux, using Context API for state management"

Accessing Memory Records

From Account Menu

1. Click your user avatar in the top-right corner
2. Select Memory... from the menu

From Dashboard User Menu

1. Open the dashboard sidebar
2. Click your profile section at the bottom
3. Select Memory Records

Managing Memory Records

Adding a Memory

1. Open Memory Records dialog
2. Click + ADD MEMORY button
3. Select memory type (Preference, Instruction, Pattern, Context, or Decision)
4. Enter a title (brief description)
5. Write the content (supports markdown and code blocks)
6. Choose status: Confirmed or Pending Review
7. Click ADD MEMORY

Editing a Memory

1. Open Memory Records dialog
2. Find the memory you want to edit
3. Click the EDIT button
4. Make your changes
5. Click UPDATE MEMORY

Memory Status

Memories have two statuses:

• Pending Review - Memories suggested by AI or recently added that need confirmation
• Confirmed - Verified memories that the AI will actively use for personalization

You can confirm pending memories by clicking the ✓ CONFIRM button, or reject them with ✗ REJECT.

Archiving Memories

To archive a memory without deleting it, click ARCHIVE. Archived memories won't be used by the AI but remain accessible if you need them later.

Best Practices

Keep Memories Concise

Aim for titles under 100 characters and content under 500 characters. Clear, concise memories are easier for the AI to understand and apply.

Use Code Blocks

For technical patterns or instructions, use markdown code blocks:

```typescript
// Preferred error handling pattern
try {
  await operation()
} catch (error) {
  logger.error('Operation failed', error)
  throw error
}
```

Be Specific

Instead of "I like clean code", write "Prefer single-responsibility functions under 20 lines".

Update Regularly

As your preferences and patterns evolve, update your memories to keep the AI's assistance relevant and helpful.

Privacy & Sharing

Memory Records are private and tied to your account. They're not shared with other users or visible in collaborative projects. They exist purely to help the AI assistant provide personalized help to you.

AI Feature Readiness

Memory Records are part of the infrastructure for upcoming AI features (chat assistant, code generation, and more). You can start building your memory library now, and the AI will use these memories when the features launch.

Searching & Filtering

Use the search bar to find specific memories by title or content.

Filter by type using the buttons at the top (Preference, Instruction, Pattern, Context, Decision) to quickly find relevant memories.

Tips

• Start with 5-10 core preferences and instructions
• Add patterns as you notice recurring choices in your code
• Document decisions when making architectural choices
• Review and confirm AI-suggested memories periodically
• Archive outdated memories instead of deleting them

Need Help?

If you have questions or feedback about Memory Records, click the chat icon in the bottom right corner to reach our support team.