Skip to main content

Mementos - Intelligent Memory System

Mementos is Bike4Mind's intelligent memory system that automatically captures, organizes, and recalls important information from your conversations. Like a personal assistant with perfect memory, it ensures that valuable insights, decisions, and context are never lost.

What are Mementos?​

Mementos are AI-curated memories extracted from your conversations. They:

  • Capture important facts, decisions, and insights automatically
  • Organize information with smart tagging and relevance scoring
  • Recall relevant memories when needed in future conversations
  • Evolve over time based on usage and importance

How Mementos Work​

Automatic Memory Creation​

When you have conversations with Bike4Mind, the AI:

  1. Identifies important information worth remembering
  2. Evaluates the significance (scored 0-10)
  3. Summarizes the key points concisely
  4. Tags the memory for easy retrieval
  5. Stores it in your personal memory bank

Memory Tiers​

Mementos are organized into three tiers based on relevance and recency:

🔥 Hot Tier​

  • Most recent and frequently accessed memories
  • Automatically injected into conversation context
  • Highest priority for recall
  • Limited capacity to maintain performance

🌡️ Warm Tier​

  • Important but less frequently accessed
  • Available for manual recall
  • Medium priority
  • Larger capacity than hot tier

❄️ Cold Tier​

  • Archived memories for long-term storage
  • Rarely accessed but preserved
  • Lowest priority
  • Unlimited capacity

Intelligent Recall​

Mementos are automatically recalled when:

  • The current conversation relates to stored information
  • You ask questions about past discussions
  • Context would be helpful for the AI's response
  • Patterns emerge across multiple conversations

Types of Information Saved​

Automatically Captured​

  • Key Decisions: "We decided to use React for the frontend"
  • Important Facts: "The project deadline is March 15th"
  • Personal Preferences: "User prefers dark mode interfaces"
  • Technical Details: "The API endpoint is https://api.example.com/v2"
  • Insights & Learnings: "Performance improved 40% after optimization"

Examples of Mementos​

[Memory] Project Architecture Decision
Context: In our discussion about the new app, we decided to use a 
microservices architecture with Docker containers. The main services 
will be: auth-service, user-service, and payment-service.
Tags: #architecture #project #microservices #docker
Score: 8.5

[Memory] User Preference - Code Style
Context: User prefers TypeScript with strict mode enabled and uses 
ESLint with Airbnb configuration. They like explicit type annotations 
even when TypeScript can infer them.
Tags: #preferences #typescript #coding-style
Score: 7.0

Managing Mementos​

Manual Controls​

While Mementos work automatically, you can also:

  • Create: Explicitly save important information
  • Edit: Update or refine existing memories
  • Delete: Remove outdated or incorrect memories
  • Search: Find specific memories by content or tags

Memory Limits​

To maintain optimal performance:

  • Hot tier: ~50 active memories
  • Warm tier: ~200 memories
  • Cold tier: Unlimited
  • Automatic tier management based on usage

Privacy & Security​

  • Mementos are private to your account
  • End-to-end encrypted storage
  • Never shared between users
  • Full export/delete capabilities

Best Practices​

Maximize Memento Value​

  1. Be Explicit: When making important decisions, state them clearly
  2. Provide Context: Include why decisions were made
  3. Use Consistent Terms: Helps with memory association
  4. Review Periodically: Check and update memories as needed

When to Create Manual Mementos​

Create manual memories for:

  • Project specifications
  • Personal preferences
  • Team decisions
  • Important URLs or credentials (encrypted)
  • Custom instructions for the AI

Memory Hygiene​

Regularly:

  • Review hot tier memories
  • Archive outdated information
  • Update changed details
  • Remove duplicate memories

Advanced Features​

Smart Tagging​

Mementos are automatically tagged with:

  • Topic categories (#project, #personal, #technical)
  • Entities mentioned (@john, @company-name)
  • Technologies (#react, #python, #aws)
  • Time references (#deadline, #milestone)

Relevance Scoring​

Each memory receives a score (0-10) based on:

  • Explicit importance markers
  • Frequency of related topics
  • User interaction patterns
  • Recency and context

Cross-Session Learning​

Mementos enable:

  • Continuity across sessions
  • Pattern recognition
  • Preference learning
  • Context accumulation

Use Cases​

Project Management​

  • Track decisions across long projects
  • Remember team member preferences
  • Maintain context between meetings
  • Document evolving requirements

Learning & Research​

  • Build knowledge incrementally
  • Connect insights across topics
  • Track learning progress
  • Create personal knowledge base

Personal Assistant​

  • Remember your preferences
  • Track important dates
  • Maintain to-do context
  • Store quick references

Technical Documentation​

  • Capture code patterns
  • Remember API details
  • Track bug solutions
  • Document system configurations

Integration with Other Features​

With Quest Master​

  • Quest results saved as memories
  • Context carried between quests
  • Learning from quest patterns

With Projects​

  • Project-specific memory contexts
  • Shared team memories (coming soon)
  • Project documentation building

With Knowledge Base​

  • Memories reference uploaded documents
  • Automatic linking between memories and files
  • Enhanced context understanding

Tips for Effective Use​

  1. Trust the System: Let automatic capture work
  2. Be Descriptive: Clear communication helps memory creation
  3. Use Natural Language: No special syntax required
  4. Iterate: Memories improve with use
  5. Provide Feedback: Correct inaccurate memories

Memory Commands​

While Mementos work automatically, you can use phrases like:

  • "Remember that..." - Explicit memory creation
  • "What do you remember about..." - Memory recall
  • "Update the memory about..." - Memory modification
  • "Forget about..." - Memory deletion
  • "Show me memories about..." - Memory search

Future Enhancements​

  • Shared Team Memories: Collaborative memory spaces
  • Memory Templates: Structured memory formats
  • External Integration: Sync with other tools
  • Advanced Analytics: Memory usage insights
  • Voice Memories: Audio note support