The Parchment throne is a text-based grand strategy game powered by multiple AI agents.
You lead a nation through diplomacy, warfare, internal politics, and economic development in a world that evolves dynamically through agent-based simulation and player interaction.
- Enable deep, emergent storytelling through AI-driven world simulation.
- Combine structured game mechanics with natural language interaction.
- Create a modular, extensible system for AI-based narrative games.
This project is powered by a multi-agent system:
-
Game Engine Agent:
Stateless, LLM-based turn processor. Queries world state via tool-calls, analyzes risks/opportunities, and triggers dynamic events. -
Conversational Agents (planned):
Persona-based temporary agents representing advisors, diplomats, generals, etc. Interact with the player and return structured, machine-readable outcomes. -
Data Backends:
- MCP (SQL): Structured, authoritative game state (nations, armies, resources, etc.)
- VectorDB (planned): Semantic embeddings of cultural, historical, and personality data
- NoSQL (planned): Dynamic documents (memories, conversations, agent state)
- ✅ Game Engine prompts for "status & analysis" phase
- ✅ Tool API design for structured world querying
- 🧪 Turn processing model with incremental queries
- 🧠 Mental energy system for conversational load (planned)
- 🧱 Modular YAML-based architecture description
- Event generation system based on context and player state
- Subchat integration with context memory and energy tracking
- Memory-aware character agents
- UI frontend (CLI or web)
ai-grand-strategy/
├── engine/ # LLM prompts, orchestration logic
├── api/ # Tool handlers (e.g. MCP SQL API)
├── data/ # Game state seeds and schemas
├── docs/ # Architecture docs, diagrams
├── LICENSE # Apache 2.0 license
├── README.md # You are here
This project is licensed under the Apache License 2.0.
See the LICENSE file for details.
Contributions, ideas, and feedback are welcome!
This project is in early stages – if you'd like to contribute, build a UI, design events, or model a new agent, open an issue or start a discussion.
Want to talk design, narrative AI, or agent-based world models?
Open a discussion or get in touch via GitHub.
This project is a creative experiment in dynamic storytelling with LLMs.
Many components are in exploration phase and subject to change.