Chat, Canvas, and nodes

A workspace combines a conversation with structured creative output. The chat and canvas are two views into the same body of campaign work.

Chat

Use chat to describe goals, add direction, answer questions, steer work in progress, and review what happens next. The interface can show tool activity, progress, approvals, and errors so important actions are not hidden behind a generic loading state.

You can cancel or steer supported runs when the current direction needs to change. Expensive generation and sensitive actions may require confirmation.

If the connection drops or the page reloads during an active run, reopen the same workspace and let the visible run state recover before sending the request again. Recovery follows the existing run identity so progress and output are not mistaken for a second request.

Canvas

Canvas is the spatial surface for arranging and reviewing work. It stores node positions and viewport state so a campaign can remain organized across sessions.

Nodes and content assets

A node is the campaign or content unit you see in the workspace. A content asset is the stored text or media behind that node. Internal compatibility code may still use the older term “artifact,” but public product language uses node and content asset.

Nodes can represent copy, images, video, captions, platform-shaped posts, or grouped campaign work. Renderer support for a platform does not automatically mean Fuzzbucket can publish directly to that platform.

Review before distribution

Check copy, media, destination, and scheduling details before publishing. Visible output is designed to remain editable and reviewable rather than disappearing into chat history.

Read the Media generation overview for generation-specific guidance.