Skip to main content
NodePad is a canvas workspace for AI chat, made by The Palaz Company. Instead of one scrolling thread, you work on an open canvas where you can run multiple AI conversations side by side, branch any message into its own thread, and merge results back together — all without losing track of where an idea came from.

​
The problem with linear chat

Every major AI assistant gives you the same interface: one message, one reply, repeat. That works for simple questions, but it breaks down when you want to explore more than one direction at once, compare answers from different models, or keep a complex project organized across dozens of exchanges. You end up with a wall of text that’s hard to navigate and impossible to restructure. NodePad replaces the scroll with a canvas. Your conversations become objects you can move, connect, and manipulate — more like a thinking tool than a chatbox.

​
Three core moves

NodePad is built around three actions that you’ll come back to constantly.
1

Chat

Open the canvas and start typing. Your first node works exactly like any AI chat — no new concepts required. Pick any supported model (Claude, GPT-5, Gemini, or open-source alternatives) and send your first message.
2

Extend to multiple chats

When a thread sparks a new question or you want to explore a different angle, spawn a second chat right on the canvas. Both conversations stay visible and spatially related, so context doesn’t get lost between browser tabs.
3

Manipulate context

Fork any message to branch the conversation from that exact point, add sticky notes to nodes to pin constraints the model should always respect, and merge threads together once you’ve reached a conclusion worth keeping.

​
What you can do in NodePad

Quick start

Send your first message and spawn your first second thread in under five minutes.

Canvas and nodes

Understand how the canvas is organized and how nodes relate to each other.

Forking and merging

Branch conversations at any message and combine threads when you’re ready.

Supported models

Compare Claude, GPT-5, Gemini, and open-weight models — switchable per message.

Sticky notes

Pin context constraints to any node so the model never forgets your rules.

Enterprise

Self-host NodePad with SSO/SCIM, RBAC, air-gapped deployment, and audit logs.
NodePad is currently in beta. Request access to try it free — see Beta access for details on how onboarding works.