OpenDCL 🧪

Build Decentraland scenes with AI. Describe what you want in plain language — OpenDCL handles the SDK7 code, ECS architecture, and deployment for you.

OpenDCL is a terminal-based AI agent that knows Decentraland inside out: every component, every pattern, every constraint. Whether you're a first-time creator or a seasoned developer, you go from idea to running scene in minutes instead of hours.

Why OpenDCL?

Building a Decentraland scene today means learning SDK7, an entity-component-system architecture, TypeScript, 60+ ECS components, parcel size limits, a QuickJS sandbox with no Node.js APIs, and a deployment pipeline. That's a steep learning curve — and it slows down even experienced developers on repetitive tasks.

OpenDCL puts an AI assistant in your terminal that already knows all of this. You describe what you want, and it writes correct, deployable scene code. It validates TypeScript as it goes, knows which 3D assets are freely available, and handles multiplayer sync, UI, animations, and deployment config without you needing to look anything up.

The result: more creators building more scenes, faster.

Use Cases

• Beginners & non-developers — "Build me a gallery with clickable paintings." Go from zero to a working scene without writing code manually.
• Experienced developers — Skip the boilerplate. Let the agent handle multiplayer sync, UI scaffolding, and deployment config while you focus on creative decisions.
• Teams & studios — Prototype scene concepts quickly and iterate before committing full dev resources.
• Land owners & brands — Build and update experiences on your LAND without hiring a dedicated development team.
• Event organizers — Spin up interactive stages, gathering areas, and event spaces fast.
• Hackathons & game jams — Go from concept to deployed scene in hours instead of days.

Features

• Branded header — on startup, displays a block-character "Decentraland" ASCII art banner with version and working directory. Falls back to a compact text header on narrow terminals
• Multi-provider LLM support — works with Claude, OpenAI, Google, Ollama (free/local), OpenRouter, and more
• Scene-aware — automatically detects your project's scene.json, SDK version, and entry points
• 20 built-in skills — scaffolding, 3D models, interactivity, UI, animations, multiplayer, authoritative server, audio/video, deployment (Genesis City & Worlds), optimization, camera control, lighting, player/avatar, NFT/blockchain, advanced rendering, advanced input, scene runtime, visual feedback, game design
• Integrated commands — /init to scaffold, /preview to launch the dev server, /tasks to manage running processes, /review to audit code
• TypeScript validation — catches type errors immediately after writing code
• Free asset catalogs — 2,700+ Creator Hub 3D models, 900+ CC0-licensed models, and 50 audio files the agent proactively suggests when building scenes
• Permission gate — prompts for confirmation before destructive bash commands, writes to sensitive files, or writes/edits outside the working directory
• Compact tool output — write shows path + size instead of file content, read shows a 5-line preview instead of 10
• Session persistence — pick up where you left off across sessions

Quick Start

# Install
npm install -g @dcl-regenesislabs/opendcl

# Run in any directory
opendcl

On first run, type /setup to configure a cloud API provider (Anthropic, OpenAI, Google, etc.).

In an Empty Folder

$ opendcl
> Create a medieval tavern scene with a bar, tables, and a fireplace

OpenDCL will scaffold scene.json, package.json, tsconfig.json, and src/index.ts with your scene.

In an Existing Scene

$ cd my-scene/
$ opendcl
> Add a click handler to the door that opens it with an animation

OpenDCL reads your scene context and modifies existing code without overwriting it.

Use with Any AI Agent

Install just the skills into your preferred AI coding agent — no opendcl installation required:

# Install all Decentraland skills (Claude Code, Cursor, Codex, Windsurf, and 35+ more)
npx skills add dcl-regenesislabs/opendcl

# List available skills first
npx skills add dcl-regenesislabs/opendcl --list

# Install specific skills only
npx skills add dcl-regenesislabs/opendcl --skill create-scene --skill multiplayer-sync

# Install globally (available in all projects)
npx skills add dcl-regenesislabs/opendcl -g

This uses the open skills CLI to copy SKILL.md files into your agent's skills directory.

Note: The full OpenDCL agent adds auto TypeScript validation, scene context detection, and slash commands (/preview, /init, /deploy) on top of these skills.

Commands

Command	Description
/setup	Configure a cloud API provider (Anthropic, OpenAI, Google, etc.)
/init	Scaffold a new Decentraland scene in the current directory
/preview	Start the Bevy-web preview server and open the scene in browser
/deploy	Deploy the scene to Genesis City or a World (auto-detects from scene.json)
/tasks	Interactively manage running background processes (preview server, etc.)
/review	Review scene code for quality, performance, and SDK7 best practices
/explain <concept>	Explain a Decentraland SDK7 concept (e.g., /explain tweens)

The agent also has a screenshot tool it can call automatically to see the running preview. On first use it asks for your permission to open a headless browser. The browser stays open for the entire session — no repeated logins.

Skills

OpenDCL loads domain-specific skills on demand based on what you're asking:

Skill	Triggered when you want to...
create-scene	Start a new project, scaffold files
add-3d-models	Load .glb models, browse free asset catalog
add-interactivity	Add click handlers, hover effects, triggers
build-ui	Create HUDs, menus, buttons with React-ECS
animations-tweens	Animate objects, play GLTF animations
multiplayer-sync	Sync state between players
authoritative-server	Server-authoritative multiplayer with anti-cheat, storage, env vars
audio-video	Add sounds, music, video screens
deploy-scene	Publish to Genesis City (LAND-based)
deploy-worlds	Publish to a World (DCL NAME or ENS domain)
optimize-scene	Fix performance, stay within limits
camera-control	Switch camera modes, cinematic cameras, cutscenes
lighting-environment	Add lights, shadows, day/night cycle, glow effects
player-avatar	Player data, emotes, attachments, NPC avatars
nft-blockchain	Display NFTs, wallet checks, smart contracts
advanced-rendering	Billboards, 3D text, materials, transparency
advanced-input	Cursor state, movement restriction, WASD patterns
scene-runtime	Async tasks, fetch, timers, realm info, restricted actions, testing
visual-feedback	Use the screenshot tool to see your scene, verify changes, iterate visually
game-design	Plan game architecture, scene limits, state management, MVP planning

How It Works

OpenDCL is built on pi-coding-agent, the agent engine behind OpenClaw. It adds Decentraland-specific:

• System prompt with SDK7 architecture knowledge (ECS, QuickJS sandbox, parcel constraints)
• Extensions that detect your project, inject context, validate TypeScript, and provide slash commands
• Skills with detailed guides for every common scene development task
• Reference docs (SDK cheat sheet, component tables, 3D asset and audio catalogs)

The agent has full access to standard coding tools (read, write, edit, bash, grep, find) and uses them to understand and modify your scene code.

Project Structure

opendcl/
├── src/
│   ├── index.ts              # CLI entry point
│   └── scene-context.ts      # Scene detection & context formatting
├── extensions/
│   ├── dcl-context.ts        # Auto-detect scene, inject metadata
│   ├── dcl-init.ts           # /init command
│   ├── dcl-preview.ts        # /preview command
│   ├── dcl-deploy.ts         # /deploy command
│   ├── dcl-setup.ts          # /setup command (cloud API provider config)
│   ├── dcl-status.ts         # Thinking/streaming status (elapsed time + tokens)
│   ├── dcl-update-check.ts   # Checks npm for newer OpenDCL versions
│   ├── dcl-validate.ts       # Post-write TypeScript validation
│   ├── dcl-screenshot.ts      # screenshot tool (headless Chrome, persistent browser)
│   ├── dcl-tasks.ts          # /tasks command (process manager)
│   ├── process-registry.ts   # Shared background process registry
│   └── permissions/           # Permission gate for dangerous operations
├── skills/                   # 20 SKILL.md files (domain expertise)
├── prompts/                  # System prompt + command templates
├── context/                  # SDK7 reference docs + asset catalog
└── tests/                    # Vitest test suites

Development

# Clone and install
git clone https://github.com/dcl-regenesislabs/opendcl.git
cd opendcl
npm install

# Build
npm run build

# Run locally
node dist/index.js

# Run tests
npm test

# Watch mode (rebuild on changes)
npm run dev

Testing

Tests are organized into two tiers:

• Unit tests (tests/unit/) — static analysis of file contents, frontmatter, and patterns
• Integration tests (tests/integration/) — dynamic imports, mock registration, cross-extension wiring, and process lifecycle verification

Integration tests catch wiring bugs (e.g., referencing a nonexistent extension file in index.ts) that static analysis alone would miss.

Adding a Skill

Skills are markdown files — no code changes needed:

1. Create skills/my-skill/SKILL.md
2. Add frontmatter with name and description
3. Write the instructions the agent should follow

---
name: my-skill
description: Brief description of when this skill should be used
---

# My Skill

Instructions for the agent...

The description field determines when the agent loads the skill. Make it specific about the user intent it covers.

Adding Context

Drop a .md file in context/ and reference it from your skills. The agent reads context files on demand to avoid bloating the context window.

LLM Providers

OpenDCL supports any provider compatible with pi-coding-agent:

Provider	API Key Env Var	Notes
Anthropic (Claude)	ANTHROPIC_API_KEY	Best quality
OpenAI	OPENAI_API_KEY	GPT-4o, o1, etc.
Google	GOOGLE_API_KEY	Gemini models
Ollama	—	Free, runs locally
OpenRouter	OPENROUTER_API_KEY	Access to many models

Set the environment variable or enter the key on first run. Switch models anytime with Ctrl+P.

Requirements

• Node.js >= 18
• npm

Contributing

Contributions are welcome! The easiest way to contribute is by adding or improving skills:

1. Fork the repository
2. Create a new skill in skills/<name>/SKILL.md
3. Test that the skill loads: npm test
4. Submit a pull request

For bugs and feature requests, please open an issue.

License

Apache-2.0