CortexGraph

About

CortexGraph addresses the critical limitation of AI assistants forgetting conversational context between sessions by implementing a sophisticated, human-like memory system. It allows AI to remember important information, preferences, and facts, which naturally decay over time unless reinforced through use, mimicking the Ebbinghaus forgetting curve. This intelligent memory management ensures that frequently used memories are strengthened and promoted to long-term storage, enabling more continuous and personalized AI interactions without the need for explicit review.

Key Features

• Knowledge graph capabilities for entities and relations.
• Smart prompting system for automatic memory saving, recall, and reinforcement during conversation.
• Temporal decay algorithm mimicking Ebbinghaus forgetting curves with reinforcement learning.
• Local, human-readable data storage ensuring privacy and transparency.
• 7 GitHub stars
• Two-layer memory architecture: JSONL for short-term and Obsidian-compatible Markdown for long-term storage.

Use Cases

• Enhancing AI assistant memory for personalized and continuous interactions.
• Preventing AI from forgetting user preferences and critical facts across sessions.
• Building AI systems with dynamic, context-aware information recall and natural forgetting mechanics.