← back
Agents on the Canvas in tldraw — Steve Ruiz, tldraw
Takeaway
Putting agents directly on a shared canvas with visible state turns AI from a sidebar assistant into a multi-agent collaborator on spatial work.
Summary
- tldraw is a whiteboard, SDK powering Replit's agent canvas, Luma AI canvas, and Stitch's annotate mode (an Angular app secretly embedding tldraw).
- Make Real (2023) was an early canvas→working-prototype demo using vision models, predating Lovable and 'vibe coding'.
- Newer canvas agents use structured-output text tools rather than diffusion models — model emits shape primitives the canvas already supports.
- Vision struggles with structured-spatial conflicts (web Y axis vs Cartesian Y axis, left/stage-left); prompt engineering carries most of the load.
- Latest demo: 'fairies' — multi-agent characters living on the canvas itself, each visible with their own thought/action state, able to collaborate (e.g., split 'draw more animals' across three fairies).
tldrawcanvas-agentsmulti-agent
Original description
At tldraw, we've been bringing agents to our infinite canvas. In December 2025, we ran a one-month experiment named Fairydraw where users could work with three fairies—virtual collaborators who work with you, with your human collaborators, and coordinate together on large tasks. Learn what we learned. Speaker info: - https://x.com/steveruizok - https://www.linkedin.com/in/steve-ruiz-61a150239/