Skip to main content
The AI chat assistant lives in your canvas sidebar and turns plain language into work on your canvas. You describe what you want, and it can plan, generate, and edit nodes for you. It is conversational by design. Ask for a single image, sketch out a whole scene, or refine what is already there, and the assistant proposes the steps before anything happens.
AI chat assistant open in the canvas sidebar

​
How a chat turns into nodes

1

Describe what you want

Type your prompt in the sidebar. Be as loose or as specific as you like, the assistant fills in the rest.
2

Review the proposed plan

The assistant replies with a plan that lays out what it intends to create or change.
3

Confirm the generation cards

Each generation shows up as a confirm card. Approve the ones you want before any credits are spent.
4

Watch nodes appear and update on your canvas

Confirmed work lands directly on your canvas as nodes, connected and ready to refine.

​
What the assistant can do

The assistant adds new nodes and edits existing ones straight from your prompt, so you can build and reshape your canvas by chatting.
Ask for an image, video, or other asset and the assistant kicks off the generation, then drops the result onto your canvas as a node.
Add an image to the chat and the assistant understands what it sees, so it can describe it, build on it, or use it as context.
Pick the model backend that fits the task. You can switch between Claude, OpenAI, and Gemini.
The assistant is metered by tokens and uses credits as it plans and generates. See credits for how the credit wallet works.
Want to know more about what the assistant creates? Explore the node types it works with and the generators behind each asset.