What a sticky note does
When you attach a sticky note to a node, NodePad includes its content as a standing instruction in the context for all subsequent messages in that thread. The model reads the sticky note alongside your messages — you don’t have to repeat the constraint yourself. The sticky note is visible on the canvas below the node it’s attached to, so you always know what’s in effect.A sticky note affects messages that come after the node it’s attached to. It does not alter any existing replies that were generated before the note was added.
What to put in a sticky note
Sticky notes work best for constraints that are stable across many turns of a conversation. One clear constraint per note keeps things legible.Tone constraints
Keep responses non-judgmental and blameless. Use neutral language throughout. Avoid hedging phrases like “it depends.”
Format requirements
Respond in bullet points only. Keep each response under 150 words. Always include a summary line at the top.
Scope limits
Only consider open-source solutions. Do not suggest changes to the database schema. Limit recommendations to the current sprint.
Role or persona
You are reviewing this as a skeptical senior engineer. Respond as a non-technical stakeholder would expect to receive this.
How to attach a sticky note
Hover the node you want to attach to
Move your cursor over the target node. The action menu appears at the edge of the node.
Select Sticky note
Click the sticky note action. An input field opens directly below the node on the canvas.
Write your constraint
Type the instruction you want to persist. Keep it direct — write it the way you’d write a system prompt, not a conversational message.
Editing and removing sticky notes
You can edit a sticky note at any time by clicking it on the canvas. Changes take effect for messages generated after the edit — replies already in the thread are not affected. To remove a sticky note, click it and select Remove. The constraint is lifted immediately; subsequent messages will no longer include it in context.Sticky notes vs. putting instructions in your message
You can always type constraints directly into a message prompt — that’s the simplest approach for a one-time instruction. Sticky notes are worth using when:- The constraint needs to hold for more than one or two replies
- You’re sharing the canvas with a collaborator and want the constraint to be visible and explicit
- You’re forking the thread and want the constraint to be clearly scoped to one branch