# Infinite canvas: the complete guide

> Updated: 2026-10-30 · Applies to KiiKii
> HTML version: https://kiikii.cc/en/doc/canvas

## Quick answer

- Node types: text, image, panorama, video, audio, generation config, and avatar.
- The top bar holds the project switcher, assets, tools, and shortcuts; the bottom holds the node toolbar and zoom controls.
- The Canvas Assistant on the right reads the current canvas and selected assets, and writes results back as new nodes.
- Select a node to inspect its generation parameters or keep generating from it.
- Only generation consumes credits; moving, connecting, and arranging nodes is free.

## 1. What the infinite canvas is
The canvas is a workbench you can zoom infinitely and arrange freely, where images, videos, audio, and text are all nodes that can be compared side by side, referenced by each other, and used to generate new content with the Canvas Assistant. Unlike the Creation Agent's linear conversation, the canvas is spatial, so you can reference, combine, and compare assets again and again—ideal for projects that orchestrate many assets.

## 2. Interface structure
The interface has four parts: the project bar at the top, the asset drawer on the left, the infinite canvas in the middle, and the assistant panel on the right. The top bar switches or renames projects and opens assets, tools, and shortcuts; the left drawer holds current assets, prompts, and the word library; the center scrolls to zoom and drags to pan; the right panel holds conversation, history, and new conversations.

## 3. Nodes and canvas operations
Nodes cover text (prompts, scripts, notes), image and panorama (reference images, storyboards, 360° scenes), video (generated shots and finished clips), audio (voiceover, music, sound effects), generation config (freezing model, aspect ratio, and duration into a node), and avatar (a digital human for talking-head content). Operations include marquee selection for bulk move or delete, undo and redo, zoom and pan, and auto layout to rearrange nodes by type when the canvas gets messy.

## 4. Assets and the asset library
The Assets drawer keeps everything on the current canvas in one place with four tabs: Current (all nodes on this canvas, counted by image and video), Assets (your account library, reusable across projects), Prompts, and Word library for templates and consistent style terminology. Every asset supports preview and Locate, which jumps to the matching node on a long canvas.

## 5. Tool panel
Once nodes pile up, the tool panel becomes the entry point for navigation and batch operations. Clicking a name locates the node; Workflows chain a series of nodes into a process you can run again; Organize / auto layout rearranges nodes by type and generation order; Export exports the canvas content or packages the selected nodes for download.

## 6. Canvas Assistant: entry points and structure
The Canvas Assistant sits on the right and is not a plain chat window: it reads the whole canvas and writes results back onto it. Chat holds the current session with result actions such as Locate video and View full size, History lists canvas conversations by time, and New chat starts a separate topic; the assistant knows which nodes exist and which asset you selected, and the input box takes natural language, with @ or a selection used to name an asset.

## 7. Canvas Assistant: how to use it
The assistant handles three requests: generate from scratch ("a romantic sunrise scene, 5 seconds"), build on the selected asset ("the animal in the image walks on the beach and sees a crab, 5-second video"), and add more rounds ("the character dances here"), carrying over the scene and style from the previous round. Write the subject, then the scene, then the shot and lighting, with duration as "5 seconds" and aspect ratio as "9:16" or "16:9"; a cancelled task produces no result and is not charged.

## 8. Selected nodes and in-node generation
Selecting a node opens a generation panel where you can add a reference image from the canvas or an upload, choose model and aspect ratio (automatic by default), and generate; results land in a new adjacent node and never overwrite the original. In-node generation is best for "one more version of this image", while the assistant is easier for combining several assets or revising continuously.

## 9. Case study 1: turning one image into a video
Add a reference image so an "Image 1" node appears, select it, and describe the shot while emphasizing that the subject must match the reference. The prompt in the case walks the animal along a wet beach, has it notice a crab and observe it, and specifies a medium tracking shot that slowly pushes in, natural daylight, and a 5-second duration. Three things raise the success rate: matching the reference, breaking the action into steps, and stating camera movement.

## 10. Case study 2: combining two images into one shot
Put two reference images on the same canvas and name both in one sentence—"Image 1 Image 2 playing together on the grass"—so the assistant uses both as references in the same video. For more control, add motion and camera: "Image 1 and Image 2 chase and play on the grass, medium tracking shot, afternoon natural light, 5 seconds"; the closer the images are in light direction and subject scale, the more natural the composite.

## 11. Three common workflows
**Image to video:** place or generate a main image, select it and write a prompt, generate a 5-second clip, locate the video to review, then move to the next shot. **Talking-head video:** place a digital human avatar node, pair it with an audio node, and generate—good for many pieces with the same person. **Multi-shot editing:** compare video nodes side by side, keep the best, tidy the order with auto layout, export, then combine into a finished piece. Turning a generation config into a reusable node keeps model, aspect ratio, and duration consistent across a batch.

## FAQ

### How is the canvas different from the Creation Agent?
The Creation Agent is linear conversation: one question, one answer, with results in time order. The canvas is spatial: images, videos, audio, and text sit on one canvas as nodes that you can reference, combine, and compare repeatedly.
### Will the Canvas Assistant delete or change my nodes automatically?
No. The assistant only adds task nodes and result nodes; existing nodes have to be deleted or moved by you. When a generation task is cancelled, no new node is produced and no credits are charged.
### How does the assistant know which image I mean?
Select a node first, then ask (for example, "the character dances here") and the assistant uses the selected asset as reference; you can also type @ to name an asset. With nothing selected, it relies on conversation context.
### Do nodes consume credits?
Only generation consumes credits; uploading, moving, connecting, and arranging cost nothing. Generation tasks run in the background and fill in the matching node when they finish.
### What if there are too many nodes to find one?
Open the Tools panel: it lists every node by name with click-to-locate. You can also use Organize / auto layout to rearrange nodes by type, then combine zoom and marquee selection for batch operations.
### Can I export works from the canvas?
A single node can be previewed, downloaded, or referenced into other features; the whole canvas can be exported through Export in the tool panel. For a finished cut, download the video nodes first and process them with an editor or Drama.