> ## Documentation Index
> Fetch the complete documentation index at: https://docs.relayhub.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Create Canvas

> Generate AI images on an infinite visual canvas

**Create** is RelayHub's image generation workspace. Instead of a chat thread, you work on an infinite canvas where prompts, reference images, and generated results are connected nodes. You can branch ideas, reuse results as references, and see exactly how each image was produced.

Open it from **Create** in the sidebar.

<Note>
  Create and Creative require the Creative Tools permission. If you do not see them in your sidebar, ask your administrator to enable creative tools for your account.
</Note>

## The Canvas

* **Toolbar** (top): editable session title, a running cost total for the session, an autosave indicator, and **+ Prompt** and **+ Reference** buttons.
* **Prompt nodes**: where you write a prompt and pick generation settings.
* **Generation nodes**: the resulting images, connected to the prompt that produced them.
* **Reference nodes**: images you upload to guide a generation.

Drag nodes anywhere. Connections between nodes show how results were produced and which references fed into them. Your work autosaves about a second after each change; the toolbar shows **Saving...** then **Saved**.

## Generating an Image

<Steps>
  <Step title="Add a prompt node">
    Click **+ Prompt** in the toolbar, then describe the image you want in the text area.
  </Step>

  <Step title="Pick a model">
    Open the **Model** dropdown. Models are grouped by provider, each with a Flash or Pro tier badge:

    | Model             | Provider | Tier  | Max resolution |
    | ----------------- | -------- | ----- | -------------- |
    | OpenAI Images 2.0 | OpenAI   | Pro   | 2K             |
    | Nano Banana       | Google   | Flash | 4K             |
    | Nano Banana Pro   | Google   | Pro   | 4K             |
    | Grok Imagine      | xAI      | Flash | 2K             |
    | Grok Imagine Pro  | xAI      | Pro   | 2K             |
  </Step>

  <Step title="Choose settings">
    * **Aspect ratio**: 1:1, 16:9, 9:16, 4:3, 3:4, 2:3, or 3:2
    * **Quality**: Low, Medium, or High (affects cost and detail)
    * **Resolution**: 1K, 2K, or 4K (options depend on the model)
  </Step>

  <Step title="Attach references (optional)">
    Click **+ Reference** to upload images (up to 25 MB each), then connect them to your prompt node. The model uses them to guide style, composition, or subject.
  </Step>

  <Step title="Click Create">
    The button shows **Working** while the image generates, typically 10 to 60 seconds depending on model and quality. The result appears as a new generation node connected to your prompt.
  </Step>
</Steps>

Each generation node shows a cost badge, and the toolbar total updates so you always know what the session has cost.

## Working with Results

Hover over a generation node and open its menu:

* **Use as Reference**: creates a new prompt node with this image attached as a reference, so you can iterate on it.
* **Branch to New Canvas**: starts a fresh canvas seeded with this image, keeping your current canvas untouched.
* **Download**: saves the full-resolution PNG.
* **Delete**: removes the generation (the session cost total adjusts).

All generations are automatically saved to your [Creative](/creative/creative-hub) gallery, so nothing is lost when you close the canvas.

<Tip>
  Use branching to explore variations without cluttering one canvas. Each branch is its own session, and the Canvas Manager in Creative keeps them all organized.
</Tip>