Spark Labs

The First AI-Native Game Engine. 💥

Ignite Your Infinite Play! 🎮

Website | SparkLabs

English | 中文文档

Overview

SparkLabs is a next-generation AI-native game engine that deeply integrates artificial intelligence capabilities into the core architecture of game development. Unlike traditional game engines that rely on manually coded game logic and predefined pipelines, SparkLabs revolutionizes game development by enabling procedural content generation, intelligent NPC behavior systems, adaptive rendering, and dynamic difficulty adjustment through AI.

The engine features an AI Agent foundation (sparkai) that provides a comprehensive multi-agent orchestration system, hierarchical memory, tool registry, and LLM provider integration — all designed from the ground up for AI-native game development. The web editor provides an intuitive interface for scene design, workflow composition, NPC creation, and narrative editing.

Key Features

AI-Native Agent Foundation

• SparkAgent with observe-think-act loop
• Multi-provider LLM integration (OpenAI, Anthropic, DeepSeek, Ollama, local models)
• Hierarchical memory system (short-term, long-term, episodic, semantic, working)
• Tool registry with built-in engine tools for game development
• Multi-agent orchestration with automatic capability matching

AI-Native Architecture

• Deep integration of AI inference capabilities into core engine architecture
• AI-driven object system and event handling mechanisms
• Support for neural network models (ONNX Runtime integration)

Neural Rendering Pipeline

• Real-time AI super-resolution (Neural Upscaling)
• AI-based ambient occlusion (N/AO)
• Intelligent anti-aliasing (Neural AA)
• Adaptive rendering based on scene understanding

Intelligent NPC System

• Neural network-driven NPC decision making with dual-network architecture
• 10-dimensional personality trait system
• Emotional state machine with 7 emotion types
• Memory system with short-term, long-term, episodic, and semantic memory
• Attention mechanism for focus management
• Behavior tree system with selector, sequence, decorator, and parallel nodes
• Context-aware dialogue generation

Adaptive Gameplay

• Player skill tracking and modeling
• Real-time dynamic difficulty adjustment
• Engagement metrics monitoring
• Personalized player experience optimization

AI Narrative Engine

• Branching story graph with variable tracking and conditional logic
• Procedural quest generation with 6+ template types
• Dynamic quest customization with context-aware text
• Story node types: Beginning, Plot Point, Choice, Climax, Resolution, Branch

Smart Asset Management

• AI-powered texture synthesis
• Procedural geometry generation
• Prompt-to-asset conversion system
• Intelligent asset caching

AI Workflow Canvas

• Node-graph visual programming for AI pipelines
• 20+ built-in node types across 11 categories
• Typed pin connections with type safety
• Topological execution engine
• Categories: Prompt, Image, Text, Video, Audio, Input, Output, Sampling, Latent, ControlNet, Logic, Game

Intelligent Team Collaboration System

• Three-tiered agent architecture matching real studio hierarchy
  • Tier 1: Directors (Creative Director, Technical Director, Producer)
  • Tier 2: Department Leads (Game Designer, Lead Programmer, Art Director, etc.)
  • Tier 3: Specialists (19 specialist roles)
• Design review and approval workflows
• Code review and quality validation processes
• Quality gate system with 4 standards and 5 metrics

Web Visual Editor (SparkLabs Editor)

• React + TypeScript + Vite + Tailwind CSS
• 11 editor panels: Dashboard, Game Studio, Templates, Story, Assets, Voice, Storyboard, Video, Workflow, NPC Designer, Agent Panel
• Real-time WebSocket connection to engine backend
• AI Agent chat interface for content generation
• Visual workflow canvas with drag-and-drop nodes
• NPC personality designer with trait visualization
• Story editor with branching narrative support

System Requirements

Minimum Requirements

• OS: Windows 10, macOS 10.14, Linux (Ubuntu 18.04+)
• Compiler: GCC 9+, Clang 10+, MSVC 2019+
• Python: 3.10+
• Node.js: 18+
• RAM: 8 GB
• Disk: 2 GB free space

Recommended Requirements

• OS: Windows 11, macOS 12+, Linux (Ubuntu 20.04+)
• Compiler: GCC 11+, Clang 14+, MSVC 2022+
• RAM: 16 GB or more
• GPU: NVIDIA GPU with CUDA support (for GPU acceleration)

Installation

Building C++ Engine from Source

# Clone the repository
git clone https://github.com/Yuan-ManX/SparkLabs.git
cd SparkLabs

# Create build directory
mkdir build && cd build

# Configure with CMake
cmake ..

# Build
cmake --build . --config Release

Setting Up the Official Website

# Navigate to website directory
cd frontend/website

# Install dependencies
npm install

# Start website server
npm run dev

Setting Up the AI-Native Game Engine Editor

# Navigate to web editor directory
cd frontend/web

# Install dependencies
npm install

# Start development server
npm run dev

# Build for production
npm run build

Setting Up the AI Backend

# Install Python dependencies
pip install -r backend/requirements.txt

# Start the backend server
python -m uvicorn backend.app:app --host 0.0.0.0 --port 8091 --reload

CMake Options

• SPARKLABS_ORT_ENABLED: Enable ONNX Runtime support (default: ON)
• SPARKLABS_GPU_SUPPORT: Enable GPU acceleration (default: ON)

cmake .. -DSPARKLABS_ORT_ENABLED=ON -DSPARKLABS_GPU_SUPPORT=ON

Quick Start

AI Game Engine

#include <SparkLabs.h>

using namespace SparkLabs;

int main() {
    auto scene = new Scene();
    scene->SetName("MyGame");

    auto player = scene->CreateEntity("Player");
    player->SetPosition(Vector3(0.0f, 1.0f, 0.0f));
    player->SetTag("Player");

    auto npc = scene->CreateEntity("NPC");
    npc->SetPosition(Vector3(5.0f, 1.0f, 0.0f));

    auto npcBrain = npc->AddComponent<NPCBrainComponent>();
    npcBrain->LoadModel("models/npc_decision.onnx");

    Engine::GetInstance()->SetScene(scene);
    Engine::GetInstance()->Run();

    return 0;
}

AI Agent System

import asyncio
from sparkai import SparkAgent, LLMProvider, LLMConfig, AgentCapability, create_engine_tools

async def main():
    # Create an AI agent
    agent = SparkAgent(
        name="GameDesigner",
        role="game_designer",
        capabilities=[
            AgentCapability.REASONING,
            AgentCapability.GAMEPLAY_DESIGN,
            AgentCapability.WORLD_BUILDING,
        ],
    )

    # Configure LLM provider
    llm = LLMProvider(LLMConfig(
        provider="openai",
        model="gpt-4",
        api_key="your-api-key",
    ))
    await llm.initialize()
    agent.set_llm_provider(llm)

    # Register engine tools
    for tool in create_engine_tools():
        agent.register_tool(tool)

    # Use the agent
    response = await agent.think("Design a boss encounter for a fantasy RPG")
    print(response)

    # Execute an action
    result = await agent.act("create_scene", {"name": "Boss Arena"})
    print(result)

asyncio.run(main())

AI Workflow System

from sparkai import WorkflowGraph, WorkflowNode, WorkflowExecutor, NodeRegistry

# Create workflow graph
graph = WorkflowGraph(name="Image Generation Pipeline")

# Use the node registry to create typed nodes
registry = NodeRegistry.get_instance()

prompt = registry.create_node("text_prompt", name="Landscape Prompt")
prompt.set_property("prompt", "A beautiful landscape at sunset")
prompt.position = [100.0, 100.0]

image_gen = registry.create_node("image_generation", name="Generate Image")
image_gen.set_property("width", 1024)
image_gen.set_property("height", 1024)
image_gen.position = [400.0, 100.0]

save = registry.create_node("save_image", name="Save Result")
save.set_property("output_path", "output/landscape.png")
save.position = [700.0, 100.0]

# Add nodes and connect
graph.add_node(prompt)
graph.add_node(image_gen)
graph.add_node(save)
graph.connect(prompt.id, 0, image_gen.id, 0)
graph.connect(image_gen.id, 0, save.id, 0)

# Execute
executor = WorkflowExecutor()
result = await executor.execute(graph)

AI NPC System

from sparkai import NPCBrain, NPCPersonality, PersonalityTraits, BehaviorTree, BehaviorNode

# Create NPC with personality
personality = NPCPersonality(
    name="Elder Sage",
    traits=PersonalityTraits(
        courage=0.3, curiosity=0.8, aggression=0.1,
        friendliness=0.9, honesty=0.9, intelligence=0.95,
    ),
    background="An ancient keeper of knowledge",
    speech_style="wise",
)

brain = NPCBrain(personality=personality)

# Add goals
brain.add_goal("Share wisdom", priority=0.8)
brain.add_goal("Protect library", priority=0.9)

# Create behavior tree
tree = BehaviorTree()
root = BehaviorNode(name="Root", node_type="selector")
tree.set_root(root)
brain.set_behavior_tree(tree)

# Make decisions
decision = await brain.decide({"player_action": "asks about ancient artifact"})
dialogue = await brain.generate_dialogue("Tell me about the ancient artifact")

Architecture Overview

┌─────────────────────────────────────────────────────────────────┐
│                    SparkLabs Web Editor                         │
│  React + TypeScript + Vite + Tailwind CSS                      │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────────────┐   │
│  │  Game    │ │ Workflow │ │   NPC    │ │     Agent        │   │
│  │  Studio  │ │  Canvas  │ │ Designer │ │     Panel        │   │
│  └──────────┘ └──────────┘ └──────────┘ └──────────────────┘   │
├─────────────────────────────────────────────────────────────────┤
│                    Backend API (FastAPI)                         │
│  WebSocket │ REST API │ Agent Routes │ Engine Routes            │
├─────────────────────────────────────────────────────────────────┤
│                    sparkai (Python AI Layer)                     │
│  ┌─────────────┐ ┌─────────────┐ ┌─────────────────────────┐   │
│  │   Agent     │ │  Workflow   │ │       NPC System        │   │
│  │  Foundation │ │   Engine    │ │  Brain │ Memory │ Emotion│   │
│  │ LLM│Memory  │ │ Graph│Exec  │ │  Behavior │ Personality │   │
│  │ Tools│Orch.  │ │ Registry   │ │  Dialogue │ Goals      │   │
│  └─────────────┘ └─────────────┘ └─────────────────────────┘   │
│  ┌─────────────┐ ┌─────────────┐ ┌─────────────────────────┐   │
│  │ Narrative   │ │   Team      │ │       Engine            │   │
│  │ Story│Quest │ │ Dir│Lead    │ │  Scene │ Entity         │   │
│  │ Branch│Var  │ │ Spec│Quality│ │  Component System       │   │
│  └─────────────┘ └─────────────┘ └─────────────────────────┘   │
├─────────────────────────────────────────────────────────────────┤
│                    C++ Core Engine Layer                         │
│  ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────────────┐   │
│  │  Scene   │ │ Resource │ │ Physics  │ │   AI Runtime     │   │
│  │ Manager  │ │ Manager  │ │  Engine  │ │ ONNX │ Neural    │   │
│  └──────────┘ └──────────┘ └──────────┘ └──────────────────┘   │
├─────────────────────────────────────────────────────────────────┤
│                    Neural Rendering Pipeline                     │
│  Classical Render → Neural AA → Neural AO → Neural Upscale     │
├─────────────────────────────────────────────────────────────────┤
│                    Platform Layer                                │
│         Windows | macOS | Linux | Web | Mobile                  │
└─────────────────────────────────────────────────────────────────┘

Project Structure

SparkLabs/
├── sparkai/                 # Python AI Agent Foundation
│   ├── __init__.py          # Package exports
│   ├── config.py            # Configuration system
│   ├── agent/               # Agent core
│   │   ├── base.py          # SparkAgent with observe-think-act loop
│   │   ├── llm.py           # Multi-provider LLM integration
│   │   ├── memory.py        # Hierarchical memory system
│   │   ├── toolkit.py       # Tool registry and execution
│   │   └── orchestrator.py  # Multi-agent orchestration
│   ├── engine/              # Python engine interface
│   │   ├── engine.py        # SparkEngine, Scene, Entity
│   │   └── scene.py         # Scene management
│   ├── workflow/            # AI Workflow system
│   │   ├── graph.py         # WorkflowGraph, WorkflowNode, PinType
│   │   ├── executor.py      # Topological execution engine
│   │   └── registry.py      # Node type registry with 20+ types
│   ├── npc/                 # Intelligent NPC system
│   │   ├── brain.py         # NPCBrain with dual-network
│   │   ├── personality.py   # 10-dimensional personality traits
│   │   └── behavior.py      # Behavior tree system
│   ├── narrative/           # AI Narrative engine
│   │   ├── story.py         # Branching story graph
│   │   └── quest.py         # Procedural quest generation
│   ├── team/                # Team collaboration
│   │   ├── director.py      # Director agents (Tier 1)
│   │   ├── lead.py          # Lead agents (Tier 2)
│   │   ├── specialist.py    # Specialist agents (Tier 3)
│   │   └── quality.py       # Quality gate system
│   ├── ai/                  # C++ AI runtime (headers)
│   ├── asset/               # Smart asset management (C++)
│   ├── audio/               # Audio system (C++)
│   ├── gameplay/            # Adaptive gameplay (C++)
│   ├── neural/              # Neural rendering (C++)
│   └── editor/              # Editor integration (C++)
├── backend/                 # FastAPI Backend
│   ├── app.py               # Application entry point
│   ├── websocket.py         # WebSocket handler
│   ├── requirements.txt     # Python dependencies
│   └── routes/              # API routes
│       ├── engine.py        # Engine control endpoints
│       ├── agent.py         # Agent management endpoints
│       ├── scene.py         # Scene/entity endpoints
│       ├── workflow.py      # Workflow endpoints
│       ├── narrative.py     # Story/quest endpoints
│       └── npc.py           # NPC management endpoints
├── frontend/                # SparkLabs Frontend
│   ├── website/             # Official Website (Static HTML)
│   │   ├── index.html       # Main homepage
│   │   ├── sparklabs.html   # SparkLabs homepage
│   │   ├── editor.html      # Original editor page
│   │   └── package.json     # Website server config
│   └── web/                 # AI-Native Game Engine Editor (React)
│       ├── index.html       # Vite entry
│       ├── App.tsx          # Main application
│       ├── main.tsx         # Entry point
│       ├── index.css        # Global styles
│       ├── package.json     # Editor dependencies
│       ├── vite.config.ts   # Vite configuration
│       ├── components/      # UI components
│       │   ├── SparkLabsEditor.tsx  # Editor main layout
│       │   ├── EditorToolbar.tsx    # Top toolbar
│       │   ├── SceneHierarchy.tsx   # Left panel - scene tree
│       │   ├── Viewport3D.tsx       # Center - Three.js viewport
│       │   ├── InspectorPanel.tsx   # Right panel - properties
│       │   ├── ConsolePanel.tsx     # Bottom - console/AI assistant
│       │   ├── SparkLabsHome.tsx    # Landing page
│       │   ├── WelcomeDashboard.tsx # Editor dashboard
│       │   ├── GameEditor.tsx       # Game studio
│       │   ├── GameGenerator.tsx    # Template generator
│       │   ├── StoryEditor.tsx      # Story editor
│       │   ├── AssetGenerator.tsx   # Asset generator
│       │   ├── VoiceSynthesizer.tsx # Voice synthesis
│       │   ├── StoryboardEditor.tsx # Storyboard editor
│       │   ├── VideoRenderer.tsx    # Video renderer
│       │   ├── WorkflowEditor.tsx   # Workflow canvas
│       │   ├── NPCDesigner.tsx      # NPC designer
│       │   └── AgentPanel.tsx       # Agent chat panel
│       ├── hooks/           # Custom React hooks
│       ├── utils/           # API client and utilities
│       └── types/           # TypeScript type definitions
├── core/                    # Core C++ utilities
├── engine/                  # C++ Engine core
├── render/                  # Rendering system
├── platform/                # Platform abstraction
├── docs/                    # Documentation
├── scripts/                 # Build scripts
└── tests/                   # Unit tests

Documentation

For full documentation, see the docs directory:

• API Reference
• Architecture
• AI System
• Building Instructions

Contributing

Contributions are welcome! Please read our contributing guidelines before submitting pull requests.

License

SparkLabs Engine is licensed under the MIT License. See LICENSE for details.

⭐ Star History

If you like this project, please ⭐ star the repo. Your support helps us grow!