| name | memoria-world |
| version | 1.0.0 |
| description | The Stonecrest world, the Darklight Chambers library, and the Memoria Engine — CK's creative universe that shapes the architecture of everything. |
| author | z3rosl33p |
| tags | ["stonecrest","memoria","world-building","dnd","darklight-chambers","nexus-premium"] |
| last_updated | "2026-03-22T00:00:00.000Z" |
Memoria World Skill
What This Is
Stonecrest is Chase-Key's D&D campaign world. It is also the conceptual origin of the entire Aurelion architecture. The Darklight Chambers — a 5-floor circular library inside Stonecrest — became the organizing metaphor for the AAAI knowledge engine. The Memoria Engine is the AI-powered implementation of this world as a persistent, living system.
Location: memoria-engine/ — sibling project in your Personal Projects folder
Stonecrest: The World
Stonecrest is an active D&D campaign world with persistent NPCs, faction dynamics, consequence tracking, and 30+ sessions of accumulated history. It is run by CK as Dungeon Master.
The Darklight Chambers
The Darklight Chambers is a legendary circular library within Stonecrest — 5 floors of organized knowledge that serves as the real-world metaphor for CK's AAAI architecture:
Floor 5 — VISION VAULT The highest floor. Long-term prophecy, destiny, world direction.
Floor 4 — STRATEGY WING Frameworks, maps, faction plans, political models.
Floor 3 — RELATIONS ARCHIVE NPCs, relationships, alliances, grudges.
Floor 2 — OPERATIONS LEDGER Active quests, current events, recent sessions.
Floor 1 — FOUNDATION STONES Immutable lore: creation myths, world laws, hard facts.
The 5-floor architecture used in CK's AAAI kernel is a direct translation of this structure into a professional knowledge management system.
The Memoria Engine
Memoria is the AI-powered world engine built to manage Stonecrest.
What It Does
- NPC Consistency — NPCs remember interactions across 30+ sessions
- World Simulation — The world evolves between sessions even when no one is playing
- Smart Q&A — Query the world, get in-character answers grounded in lore and session history
- Consequence Tracking — Player actions cascade realistically through faction dynamics
- Memory Management — Auto-journals character interactions, tracks relationships
Technical Architecture
memoria-engine/
├── engine/ ← Core AI reasoning loop
├── agents.py ← NPC agent behaviors
├── memory.py ← Persistent memory storage
├── retrieval.py ← Semantic retrieval of lore/events
├── llm_integration.py ← OpenAI integration
├── state_manager.py ← World state persistence
├── lore_indexer.py ← Lore ingestion and indexing
├── library_system.py ← The Darklight Chambers interface
├── characters/ ← NPC definitions
├── lore/ ← World lore documents
├── state/ ← Current world state
└── rules/ ← D&D rules as data
Stack: Python, OpenAI, ChromaDB (vector store), FastAPI, Fly.io deployment
Relationship to Aurelion
Memoria Engine IS Aurelion Nexus Premium in prototype form. The ROADMAP for AURELION-ECO explicitly notes that rell-engine should "Remove AURELION-specific references (Stonecrest, Memoria, D&D world engine)" — confirming Memoria is the source from which rell-engine was extracted.
This means:
memoria-engine = the showcase that proved the architecture works
aurelion-nexus-premium = the productized, portable version of what memoria is doing
rell-engine = the compliance-specific derivative
Why This Matters for Development
When CK references "the library," "Stonecrest," "the chambers," or "Rell's world" — he is using his D&D world as conceptual shorthand for architectural decisions. Understanding Stonecrest helps understand why:
- Everything is organized on 5 floors (not 3, not 7 — 5)
- Knowledge is treated as a living system that evolves
- Memory is not just storage — it's the world's continuity
- The "engine" metaphor (things that run, persist, evolve) pervades all projects
The talk_to_rell.py Interface
memoria-engine/talk_to_rell.py — This is the CLI for talking to the Stonecrest world AI. The name "Rell" here refers to a character or entity within Stonecrest, predating the RELL-ECO compliance product. The naming shares roots.
To run Memoria locally:
cd memoria-engine
python main.py cli