The tldraw SDK is an infinite canvas engine for React, used as the foundation for whiteboards, diagramming tools, and other canvas apps. It ships a default whiteboarding tool set and the primitives to build new shapes, tools, bindings, and UI components, and its Editor API drives the canvas contents, effects, and behaviors at runtime.
Built-in behavior covers what a canvas app expects: copy and paste, undo and redo, cross-tab sync, selection and transformation with nested transforms, hit-testing through a geometry system, alignment, distribution, reordering, and drag and drop. State lives in a signals library and record store with events and side effects. The canvas renders through the DOM, so it can embed anything the browser supports, including websites from YouTube, Figma, and GitHub, and it works across desktop, touch screens, tablets, and mobile.
For agent builders, tldraw provides canvas primitives for building with LLMs and an Agent starter kit in which AI agents read, interpret, and modify canvas content. Other starter kits cover multiplayer collaboration, a drag-and-drop workflow node builder, canvas-based AI chat, branching chat, an image generation pipeline builder, and WebGL shaders. Documentation is also published as an LLMs.txt file, and packages include versioned docs and release notes for coding agents.
Features
- Multiplayer: self-hostable real-time collaboration with live cursors, viewport following, and cursor chat
- Drawing and diagramming: pressure-sensitive drawing, geometric shapes, rich text, arrows, snapping, images, video, and image export
- Runtime API: the Editor API gives programmatic control of canvas contents and behavior
- Extensibility: custom shapes, tools, bindings, UI components, themes, side effects, and event hooks
- AI integrations: canvas primitives for building with LLMs
- Agent starter kit: AI agents that read, interpret, and modify canvas content
- Other starter kits: multiplayer, workflow node builder, canvas chat, branching chat, image pipeline, and shader kits
- DOM canvas: embeds web content such as YouTube, Figma, and GitHub
- Accessibility: dark mode, motion controls, screen reader support, and keyboard navigation
