para-memory-files

PARA knowledge management organizes durable long-term memory by storing atomic facts as YAML files within structured Project, Area, and Resource folders. This prevents important context from getting lost across sessions by giving an AI a durable, organized memory.

The best way to recall prior context is using semantic or keyword search (qmd) over your indexed personal folder. The system prioritizes retrieval by maintaining a summary.md file and applying memory decay tiers to surface relevant daily notes.

You save and update knowledge by writing durable details directly into items.yaml files and updating relevant summaries automatically. This ensures your AI maintains continuity by capturing the raw event stream in timestamped markdown files.

Yes, file-based memory supports planning by maintaining separate plans at the project root. It runs weekly synthesis of summaries and updates summary.md files to ensure retrieval prioritization with memory decay tiers keeps your knowledge current.

You need to archive inactive items in a PARA knowledge graph when details are no longer actively relevant to your Projects, Areas, or Resources. This maintains a clean structure and ensures your AI retrieves only current, actionable context.