io.github.tldraw/tldraw MCP Server
io.github.tldraw/tldraw
Draw and visually collaborate with AI agents on tldraw's infinite canvas.
What is the io.github.tldraw/tldraw MCP server?
The tldraw MCP server enables AI agents to draw, create, and visually collaborate on tldraw's infinite canvas. It provides canvas primitives and runtime APIs for building AI-integrated whiteboarding and diagramming experiences.
This server connects AI agents to tldraw's feature-complete canvas engine, allowing agents to create shapes, drawings, diagrams, and interactive visual content. Use it to enable visual collaboration between users and AI, build canvas-based workflows, or let agents interpret and modify canvas content in real-time.
How to install io.github.tldraw/tldraw
Copy-paste configuration for popular MCP clients.
Tools & capabilities
Tools this server exposes to the agent.
Canvas drawing and creation— Create and manipulate shapes, drawings, geometric shapes, text, arrows, and images on the canvasRuntime API— Drive the canvas at runtime with the Editor API to programmatically control canvas state and interactionsShape and tool extensibility— Define custom shapes, tools, bindings, and UI components for specialized canvas applicationsMultiplayer collaboration— Real-time collaborative editing powered by @tldraw/sync with self-hostable infrastructureExport and rendering— Export canvas content as images and leverage DOM canvas rendering for embedded content like YouTube, Figma, and GitHub
Use cases
- Enable AI agents to sketch, annotate, and mark up diagrams alongside conversations
- Build visual programming or workflow automation tools with drag-and-drop node builders
- Create AI-powered whiteboarding apps where agents interpret and modify canvas content
- Generate and visualize image pipelines or shader-based visual effects
- Implement collaborative canvas experiences where users and agents co-create visual content
io.github.tldraw/tldraw MCP server FAQ
It's an MCP server that connects AI agents to tldraw's infinite canvas engine, enabling visual collaboration, drawing, diagramming, and real-time canvas manipulation.
The tldraw SDK is free for development. Production use requires a license key; visit tldraw.dev/pricing for details.
Use the remote connection methods: streamable-http at https://tldraw-mcp-app.tldraw.workers.dev/mcp or SSE at https://tldraw-mcp-app.tldraw.workers.dev/sse. Configure these endpoints in your MCP client settings.
Yes, tldraw includes self-hostable real-time collaboration via @tldraw/sync, powered by Cloudflare Durable Objects.
Yes, tldraw is fully extensible. You can create custom shapes, tools, bindings, UI components, and event hooks for specialized applications.
Agents can create and modify shapes, drawings, text, arrows, and images; read and interpret canvas content; drive the canvas via the Editor API; and participate in real-time collaborative sessions.
README (reference)
Source of truth, from the repository.
Feature highlights
tldraw provides a feature-complete infinite canvas engine designed to be the foundation for any canvas app. Create custom shapes, tools, bindings and UI components for a custom experience. Use the default whiteboarding tool set or use the library's primitives to build entirely new shapes and interactions.
- Multiplayer — self-hostable real-time collaboration with
@tldraw/sync - Drawing and diagramming — pressure-sensitive drawing, geometric shapes, rich text, arrows, snapping to shapes, edge scrolling, image and video support, image export
- Runtime API - drive the canvas at runtime with the Editor API
- Fully extensible — custom shapes, tools, bindings, UI components, side effects, and event hooks
- AI integrations — canvas primitives for building with LLMs
- DOM canvas — web rendering supports anything the browser supports, including embedded websites from YouTube, Figma, GitHub, and more
- Broad support — works in any browser across desktop, touch screens, tablets, and mobile devices
Who's using tldraw
The tldraw SDK powers canvas experiences in products from Google, Shopify, BlackRock, Autodesk, ClickUp, Replit, Google Stitch, Luma, Runway, Padlet, Mobbin, Jam, Craft, Honeycomb, SchoolAI, Brisk, CADChat, bigpi, Genio, Pollination, Legendkeeper, Matilda, Aries, Alai, Dirac, and many more.
See our showcase for case studies on how teams build with tldraw.
Quick start
Install the tldraw package:
npm i tldraw
Then, use the <Tldraw /> component in your React app:
import { Tldraw } from 'tldraw'
import 'tldraw/tldraw.css'
export default function App() {
return (
<div style={{ position: 'fixed', inset: 0 }}>
<Tldraw />
</div>
)
}
Starter kits
Starter kits provide the custom shapes, tools, and user interface needed for common applications. Each kit is MIT-licensed. Hack together a prototype, build out an app on top, or reference the code in a larger project.
Start building with:
npx create-tldraw@latest
- Multiplayer — self-hosted real-time collaboration powered by
@tldraw/syncand Cloudflare Durable Objects, the same stack behind tldraw.com - Agent — AI agents that read, interpret, and modify canvas content
- Workflow — drag-and-drop node builder for automation pipelines, visual programming, and no-code platforms
- Chat — canvas-powered AI chat where users sketch, annotate, and mark up images alongside conversations
- Image pipeline — node-based builder for image generation pipelines
- Branching chat — AI chat with visual branching, letting users explore and compare different conversation paths
- Shader — WebGL shaders that respond to canvas interactions
Local development
The development server runs the examples app at localhost:5420. You'll need Node.js ^20.0.0. Clone the repo, then enable corepack for the correct yarn version:
npm i -g corepack
Install dependencies and start the dev server:
yarn
yarn dev
Documentation
Documentation for the most recent release can be found on tldraw.dev/docs, including reference docs. Our release notes can be found here.
For more agent-friendly docs, see our LLMs.txt.
From 5.1.x onward, published packages also include a DOCS.md file with the appropriate docs-site content and a generated RELEASE_NOTES.md file with versioned release notes. If you or your coding agent need release notes for your installed version specifically, check these files alongside the source in your node_modules folder.
Community
- Discord — questions, feedback, and discussion
- Twitter/X — news and updates
- Submit an issue — bug reports and feature requests
Contributing
We are not accepting contributions at this time. If you've found a bug or have a feature request, please create an issue and we can discuss it there. See our contributing guide for details.
License
The tldraw SDK is provided under the tldraw license. You can use the SDK freely in development. Production use requires a license key. Visit tldraw.dev to learn more.
Trademarks
Copyright (c) 2024-present tldraw Inc. The tldraw name and logo are trademarks of tldraw.
Please see our trademark guidelines for info on acceptable usage.
Contributors
<a href="https://github.com/tldraw/tldraw/graphs/contributors"> <img src="https://contrib.rocks/image?repo=tldraw/tldraw&max=400&columns=20" width="100%"/> </a>Star history
<a href="https://star-history.com/#tldraw/tldraw"> <picture> <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=tldraw/tldraw&type=Date&theme=dark" /> <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=tldraw/tldraw&type=Date" /> <img src="https://api.star-history.com/svg?repos=tldraw/tldraw&type=Date" alt="Star History Chart" width="100%" /> </picture> </a>Related MCP servers
Crypto trading intelligence MCP — 34+ endpoints, x402 pay-per-use, AI agent strategy & execution
View repository →
NeuroDock Chronometric
Time-of-day, session boundaries, and break prompts for time-blind neurodivergent users working with Claude.

Local typed-edge cognitive graph that externalizes memory of people, projects, and decisions for neurodivergent professionals.

NeuroDock Guardrail
Advisory detectors for rumination, hyperfocus, and sycophancy—never silently blocks.

Decompose vague goals into atomic 5-90 minute tasks and surface the next safe action.

NeuroDock Translation
Decode corporate ambiguity and rewrite messages with AI-powered translation for neurodivergent professionals.
