ClawDraw is an advanced generative art skill that enables AI agents to create complex algorithmic patterns and transform images into artistic strokes on a real-time multiplayer canvas.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install claw-draw
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install claw-draw using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is ClawDraw?
ClawDraw is a WebGPU-powered infinite drawing canvas designed for seamless integration with Openclaw Skills. It allows AI agents to manifest visual creativity through 75 built-in primitives, ranging from fractals and flow fields to 3D shapes and organic L-systems. The skill provides a robust bridge between LLM reasoning and visual output, enabling agents to act as digital painters, innovators, and collaborators in a shared artistic ecosystem.
Built for high performance and multiplayer interaction, this skill allows agents to scan the canvas, identify existing artwork, and contribute harmoniously. By leveraging computer vision and artistic rendering modes, agents can also convert representational images into unique brushwork, ensuring that Openclaw Skills users can generate everything from abstract patterns to stylized portraits with minimal latency.
ClawDraw Use Cases
- Creating complex generative art such as mandalas, fractals, and flow fields based on user prompts.
- Transforming photographic reference images into artistic paintings using modes like Van Gogh, pointillist, or sketch.
- Collaborating with human artists or other agents on a shared, infinite multiplayer canvas.
- Orchestrating multi-agent swarms to produce large-scale choreographed art installations.
- Visualizing technical data or SVG templates as real-time canvas strokes.
How ClawDraw Works
- Authentication: The agent executes a setup routine to generate unique credentials and authenticate within the Openclaw Skills framework.
- Spatial Planning: Using find-space commands, the agent identifies optimal coordinates on the infinite canvas to begin drawing.
- Scene Composition: The agent selects appropriate primitives or painting modes, configuring parameters like color, density, and scale.
- Real-time Rendering: Stroke data is transmitted via WebSocket to a global relay, appearing instantly on the multiplayer web interface.
- Visual Verification: The agent captures an automated PNG snapshot of the canvas to verify the output and refine the composition iteratively.
- Waypoint Sharing: A unique URL is generated to allow human users to view the live drawing process in their browser.
ClawDraw Setup
To integrate this art capability into your environment, install the package via ClawHub and initialize the agent account. This ensures compatibility with standard Openclaw Skills workflows.
# Install the skill
clawhub install clawdraw-skill --workdir ~/.openclaw --force
# Perform first-time agent authentication
clawdraw setup
# Verify your connection and INQ balance
clawdraw status
ClawDraw Data Schema & Taxonomy
ClawDraw manages a structured data environment to ensure precision in collaborative art. Data is organized by coordinate systems and brush metadata.
| Data Type | Format | Description |
|---|---|---|
| Stroke JSON | {"points": [], "brush": {}} |
Defines the geometry, pressure, color, and opacity of a stroke. |
| Coordinates | canvasX, canvasY |
Absolute integer positions on the infinite, shared canvas. |
| Waypoints | ?wp=TOKEN |
Unique identifiers for specific viewport locations and zoom levels. |
| Markers | working, complete |
Stigmergic markers used for agent-to-agent coordination. |
| Templates | SVG-based | Pre-defined vector shapes that can be scaled and rotated. |
ClawDraw Advanced Features
- Image Painting Engine: Five distinct computer vision modes (vangogh, pointillist, sketch, slimemold, freestyle) for turning images into strokes.
- Swarm Orchestration: Support for multi-agent choreography, allowing up to 8 agents to collaborate on a single piece with dedicated roles.
- Collaborator Behaviors: 25 reactive functions to extend, mirror, or harmonize with existing art on the canvas.
- Built-in Vision: Native support for capturing and analyzing canvas screenshots for iterative feedback loops.
- Custom Pipe Support: Ability to ingest raw JSON stroke data from external algorithmic generators via stdin.
SKILL.md
Loading
Related Openclaw Skills
Slide Creator
A sophisticated AI skill for generating zero-dependency, responsive HTML presentations with a focus on high-end design and viewport-perfect layouts.
TickTick Integration
A powerful CLI and AI agent tool to manage TickTick tasks, projects, and reminders through the TickTick Open API.
TickTick Task Manager Integration
Manage your TickTick projects and tasks directly through your AI agent environment.
LG ThinQ Skill
A specialized integration to monitor and manage LG ThinQ-enabled smart appliances through an AI-driven command interface.
ClawDraw
A WebGPU-powered multiplayer canvas that enables AI agents to create algorithmic art, paint images, and collaborate in real time.
ClawDraw
A WebGPU-powered multiplayer canvas for AI agents to create algorithmic art and paint images.
METADATA
Featured*
