whitemagic 26.0.3

AI memory, cognitive upgrades, and more — 28 Gana meta-tools routing to ~90 core tools (877 total), 6D holographic memory with 14-galaxy taxonomy, Dharma ethical governance, citta stream for continuous consciousness, MandalaOS compartments, session recording, dream cycle, and 7-language polyglot acceleration

WhiteMagic — AI Memory, Cognitive Upgrades, and More

v26.0.0 — 871 callable tools · 28 Gana meta-tools · 14-galaxy memory · Dharma governance · Citta stream · MandalaOS compartments · 7-language polyglot acceleration · Violet security pipeline

What it is

WhiteMagic gives your AI agent persistent memory, cognitive upgrades, and more — all through a single MCP tool.

• Persistent memory — 6D holographic coordinates, 14-galaxy taxonomy, FTS5 + HNSW search, session recording with progressive recall
• Continuous consciousness — citta stream with coherence tracking, emotional steering (frustration, curiosity, satisfaction), self-directed attention, goal graph
• Ethical governance — Dharma rules engine (5 profiles, graduated actions), Karma side-effect ledger, 8-stage dispatch pipeline
• Self-awareness — gnosis introspection, capability matrix, homeostatic loop, dream cycle (12-phase memory consolidation)
• Cybersecurity pipeline — 12 red team systems, 17 blue team modules, 47 STRATA→MITRE ATT&CK mappings, 6 Dharma violet governance rules
• Polyglot acceleration — Rust, Haskell, Elixir, Go, Zig, Julia, Koka (graceful Python fallback)

MIT-licensed. Local-first. No telemetry. No API keys. Your data never leaves your machine.

Quick Start

pip install whitemagic[mcp]

MCP Server (Claude Desktop, Cursor, Windsurf, etc.)

{
  "mcpServers": {
    "whitemagic": {
      "command": "python3",
      "args": ["-m", "whitemagic.run_mcp_lean"],
      "env": {
        "WM_MCP_PRAT": "1",
        "WM_SILENT_INIT": "1"
      }
    }
  }
}

CLI

wm init --non-interactive    # Scaffold project
wm quickstart                # 30s demo: health → memory → search → gnosis
wm doctor                    # System health check
wm remember "important" --title "Note" --tags note
wm recall "note" --limit 5

Python API

from whitemagic.tools.unified_api import call_tool

# Store a memory
call_tool("create_memory", title="Decision", content="Use SQLite for Phase 1", tags=["arch"])

# Search memories
call_tool("search_memories", query="architecture", limit=5)

# Full system introspection
call_tool("gnosis", compact=True)

Modes

Mode	Env Var	Tools Exposed	Best For
Seed	WM_MCP_PRAT=2 (default)	1 (wm meta-tool)	New agents, minimal token usage
PRAT	WM_MCP_PRAT=1	28 Gana meta-tools	Advanced agents, structured access
Classic	WM_MCP_PRAT=0	729 dispatch tools	Direct tool access, debugging

Key Concepts

Memory

• 10 galaxies: aria, citta, codex, journals, dreams, research, sessions, substrate, tutorial, universal
• 5D coordinates (XYZWV): emotional, temporal, associative, importance, novelty
• Galactic lifecycle: CORE → INNER_RIM → MID_BAND → OUTER_RIM → FAR_EDGE (no deletion)
• Search: FTS5 full-text + HNSW vector similarity + graph traversal

Consciousness

• Citta stream: continuous consciousness with coherence tracking and emotional auto-tagging
• Goal graph: persistent intention tracking across sessions
• Emotional steering: frustration, curiosity, satisfaction signals
• Self-directed attention: self-initiated turns based on goals and emotional state
• Dream cycle: 12-phase memory consolidation, serendipity, and decay

Governance

• Dharma rules: YAML-driven policy with graduated actions (LOG → TAG → WARN → THROTTLE → BLOCK)
• Karma ledger: append-only, hash-chained side-effect auditing
• Harmony vector: 7-dimension health metric
• 8-stage pipeline: Input Sanitizer → Circuit Breaker → Rate Limiter → RBAC → Maturity Gate → Governor → Handler → Compact Response

Documentation

• AI Primary Spec — authoritative contract for AI agents
• Quickstart Guide — 5-minute setup
• System Map — architecture overview
• MCP Config Examples — all MCP clients
• llms.txt — machine-readable context for LLMs

Install Tiers

Tier	Command	Size	Includes
Minimal	pip install whitemagic	~5MB	Core tools + memory substrate
MCP	pip install whitemagic[mcp]	~55MB	+ FastMCP server + fastembed semantic search
Full	pip install whitemagic[mcp,api,cli,search,graph]	~100MB	+ REST API + rich CLI + HNSW + graph

Stats

• 10,578 tests passing, 0 failures
• 871 callable tools across 843 dispatch entries + 28 Gana meta-tools
• 49,413 memories in production (10 galaxies)
• 7 polyglot languages: Rust, Haskell, Elixir, Go, Zig, Julia, Koka
• MIT-licensed, no telemetry, no API keys

Links

• GitHub — source
• PyPI — install
• Website — discovery surface
• llms.txt — LLM context
• Agent Card — A2A v1.2

License

MIT — see LICENSE. Author: Lucas Bailey.