Memory

What it does

The Memory skill enables AI agents to store and retrieve information across multiple sessions, providing a persistent knowledge base. It allows for writing notes with titles, content in markdown format, tags, folders, and project associations. You can then search these notes using semantic search, filter by title, date or permalink, and build context around specific topics or URLs. This facilitates richer conversations and improved long-term understanding by leveraging past interactions.

When to use it

• Saving learning resources: Capture useful information discovered while researching a topic (e.g., TypeScript utility types).
• Documenting architectural decisions: Record important choices made during development, including rationale and alternatives considered.
• Organizing meeting notes: Store and easily find notes related to specific meetings or projects.
• Retrieving previous work: Quickly access context from past conversations or tasks using the "continue-conversation" feature.
• Building a personal knowledge base: Create a searchable repository of notes, tagged for easy organization and retrieval.

Key capabilities

• Note Creation: Write notes with titles, markdown content, tags, folder assignments, and project associations.
• Semantic Search: Search notes using natural language queries.
• Filtering & Sorting: Filter search results by title, date (within a timeframe), or permalink.
• Context Building: Generate related notes based on a URL or topic.
• Recent Activity Tracking: View recent note creation and modification activity.
• Conversation Continuation: Retrieve prompts to continue previous work based on context.

Example prompts

• "Write a note titled 'React Hooks' with the content '# React Hooks\n\n- useState, useEffect, useMemo'"
• "Search for notes related to 'authentication' within the 'myproject' project."
• "Build context around memory://topic/api"

Tips & gotchas

• Prerequisites: The basic-memory package must be installed via pip: pip install basic-memory.
• Note Format: Notes are stored in markdown format with YAML frontmatter for metadata (title, tags, created date).
• Project Association: Notes can be associated with specific projects to keep related information organized.