Skip to main content
A node is the smallest piece of content on the canvas: one message you send, or one reply the model generates. Nodes link together into threads, threads live on the canvas, and every action that gives NodePad its power — forking a branch, attaching a sticky note, merging results — begins by pointing at a specific node.

​
Anatomy of a node

Every node on the canvas has the same basic structure regardless of which model produced it or where it sits in a thread.
The text of the message or reply, rendered with full Markdown support. Long replies collapse to a preview with an option to expand, so the canvas doesn’t become unwieldy.
A label showing which model generated the reply — Claude, GPT-5, Gemini, or an open-source model. You can change the model for the next reply at any time.
Hover a node to reveal its action menu: fork, attach a sticky note, copy, or start a merge from this point. Everything you can do with a node is one click away.
Any sticky notes attached to a node are displayed inline below the node content. They persist visually so you always know what constraints are active.

​
How nodes form threads

Nodes connect sequentially: your message, then the model’s reply, then your next message, and so on. That chain of connected nodes is a thread. Threads appear as vertical columns on the canvas, and you can have as many threads as your work demands. When you fork a node, a new thread branches off from that point. The forked thread inherits all the context from every node above the fork — it just grows in a new direction from there.
The node you fork from stays part of the original thread. Forking never removes or alters existing nodes — it only creates new ones.

​
What you can do from any node

Fork

Create a new branch that starts with the same context as the current thread up to this point.

Attach a sticky note

Pin a constraint — tone, format, scope — that persists for all subsequent messages in this thread.

Merge

Begin a merge from this node, bringing a result from another thread into a new synthesis node.

​
Nodes are non-destructive

No action you take on a node — forking, merging, attaching a sticky note — modifies the node itself. Your conversation history is always intact. You can revisit any node in any thread at any time and continue from that point.
If you want to try a completely different approach to a question without losing your current progress, fork from the node immediately before your last prompt. The fork starts fresh from that context without touching the original thread.