---
title: "Game Examples"
description: "AI-powered and agentic games with elizaOS (TypeScript)"
---

Game examples live under `packages/examples/`. They use `AgentRuntime`, plugins, and (for LLM-backed demos) `ModelType` for decisions.

## Text adventure (LLM)

Dungeon crawl with AI-chosen actions.

| Entry                         | Notes                          |
| ----------------------------- | ------------------------------ |
| `examples/text-adventure/game.ts` | Set `OPENAI_API_KEY`; use `LOG_LEVEL=fatal` for quiet logs |

```bash
cd examples/text-adventure
cp .env.example .env   # add OPENAI_API_KEY
bun install
LOG_LEVEL=fatal bun run game.ts
```

---

## Tic-tac-toe (no LLM)

Minimax via custom model handlers — no API calls.

```bash
bun run examples/tic-tac-toe/game.ts
```

---

## Game of life (no LLM)

Multi-agent simulation with algorithmic decisions.

```bash
bun run examples/game-of-life/game.ts
```

---

## Core pattern (text adventure)

```typescript
import { AgentRuntime, ModelType } from "@elizaos/core";
import { openaiPlugin } from "@elizaos/plugin-openai";
import { plugin as sqlPlugin } from "@elizaos/plugin-sql";

const runtime = new AgentRuntime({
  character: {
    name: "Adventurer",
    bio: "Exploring a dangerous dungeon.",
    system: `You are playing a text adventure. Choose ONE action: move north, take sword, attack, …
Respond with ONLY the command.`,
  },
  plugins: [sqlPlugin, openaiPlugin],
});

await runtime.initialize();

const decision = await runtime.useModel(ModelType.TEXT_SMALL, {
  prompt: "You are in the entrance. Exits: north. What do you do?",
  maxTokens: 50,
  temperature: 0.3,
});

console.log(String(decision));

await runtime.stop();
```

---

## Next Steps

<CardGroup cols={2}>
  <Card title="Examples Overview" icon="grid-2" href="/examples/overview">
    All examples
  </Card>
  <Card title="Create a Plugin" icon="puzzle-piece" href="/guides/create-a-plugin">
    Extend the runtime for your game
  </Card>
</CardGroup>