cortex-visualize

Launch the interactive unified neural graph visualization. Use when the user says 'show visualization', 'show me the graph', 'visualize memories', 'show memory map', 'open neural graph', or when a visual overview of the memory system or cognitive profile would be helpful.

Visualize — Interactive Neural Graph

Keywords

visualize, graph, neural graph, memory map, show memories, visual overview, entity graph, methodology graph, show profile, interactive

Overview

Launch the interactive browser-based unified neural graph combining methodology profiles, memories, and knowledge graph connections. Features 2D force-directed layout with domain separation, emotional tagging, quality scoring, and global memory visualization.

Use this skill when: The user wants a visual overview, is exploring the knowledge graph, or needs to present/screenshot Cortex's state.

Workflow

Launch Unified Neural Graph

cortex:open_visualization({})

Or filter to a specific domain:

cortex:open_visualization({
  "domain": "cortex"
})

Opens in the browser at http://127.0.0.1:3458. Features:

  • 2D force-directed graph with methodology nodes, memory nodes, and entity nodes
  • Color coding by domain, heat level, memory type, and emotional state
  • Global memories shown in pink (#FF4081) with double-ring, connected to all domain hubs
  • Quality scoring with colored arcs (green/amber/red)
  • Interactive -- click nodes for details, drag to explore, scroll to zoom
  • Filters -- All, Methodology, Memories, Knowledge, Emotional, Protected, Hot, Global
  • Auto-shutdown after 10 minutes idle

Get Graph Data (Programmatic)

For custom visualization or analysis:

cortex:get_methodology_graph({
  "domain": "<optional filter>"
})

Returns raw graph data (nodes + edges) that can be processed or exported.

Tips

  • Server auto-shuts down after 10 minutes of no browser activity
  • Bookmark the URL: http://127.0.0.1:3458 stays consistent across launches
  • Screenshots: The Cyber Obsidian theme is designed to look great in dark mode
  • Large datasets: The graph handles 1000+ nodes with force-directed layout