Why a canvas instead of a chat window
Most AI apps give you one thread: a vertical scroll of messages from top to bottom. That works fine for a single question, but it breaks down when your work spreads across multiple topics, approaches, or models. You end up juggling browser tabs, losing context between sessions, or — worse — letting a promising tangent die because there was no clean place to put it. The canvas solves this by adding a second dimension. Instead of a single line of messages, you get a surface where threads can be neighbors. A research session might have three threads running in parallel. An engineering decision might have a main thread with two forks for second opinions. You see all of it at once, spatially arranged the way the problem actually fits in your head.What lives on the canvas
Chat threads
Each thread is a sequence of nodes — messages and replies — organized vertically. Threads are the primary unit of work on the canvas.
Forked branches
When you fork a node, a new thread grows from that point. The branch appears on the canvas beside its parent, sharing the same context up to the fork.
Synthesis nodes
When you merge two threads, a synthesis node captures the reconciled result. It lives on the canvas alongside the threads that produced it.
Sticky notes
Sticky notes attach directly to nodes. They’re visible on the canvas as a persistent reminder of the constraints in effect for that thread.
Navigating the canvas
Pan to explore
Click and drag any empty area of the canvas to move around. Threads don’t snap to a grid — place them wherever the layout makes sense to you.
Zoom to focus
Scroll or pinch to zoom in on a specific thread or zoom out to see your full workspace at once.
Canvases and sessions
Each canvas represents a working session or project. You can create multiple canvases — one per project, client, or research question — and switch between them without losing the spatial layout you’ve built up.NodePad is currently in beta. Join the waitlist to get access.