Create Image node: generate and edit images in Workflows
Create Image node
Create Image is the one node for making and changing images in a workflow. Start with a prompt to generate a new image, or connect an image to edit it. Add more images as references to guide the result. It covers most creative steps in a workflow: new backgrounds, new angles, styling a garment on a model, or creating a scene from scratch.
How it works
- Generate: leave the image input empty and describe what you want. Create Image makes a new image from your prompt.
- Edit: connect an image and describe the change. The first connected image is the one that gets edited.
- References: connect more images after the first one. They guide the result (a colorway, a fabric, a model, a pose or a scene) without being edited themselves.
The prompt box reminds you of this: "Describe a new image, or the change to make to Image 1".
Inputs and settings
- Images: up to 6 in total. Image 1 is edited; the rest are references.
- Prompt: what to create, or what to change in Image 1. You can tag connected images in the prompt, for example $product.
- Engine: Fast (default), Pro or Max.
- Resolution: 1K (default), 2K or 4K. Max renders at up to 2K.
- Aspect ratio: pick a ratio such as 1:1 or 4:5, or choose Auto to keep the shape of the first connected image. With no image connected, Auto falls back to square (1:1).
When Create Image is connected to a Batch Input, it runs once for each image in the batch, applying the same prompt and settings to every one.
Choosing an engine
The engine you choose affects generation speed, image quality, prompt adherence, and credit usage. Use Fast for quick iterations, Pro for a balance of quality and speed, and Max when your image needs greater detail or more complex edits.
Engine | Best for | What to expect |
|---|---|---|
Fast | Brainstorming, testing prompts, and simple edits | The fastest and most affordable option. Quality and prompt adherence may be lower than Pro or Max. |
Pro | Everyday image generation and editing | A balance of quality and generation time, with better detail and prompt adherence than Fast. |
Max | Detailed prints, textures, complex edits, and reference-based changes | The highest-quality option, especially for fine details and complex instructions. It takes longer and costs more credits. |
Credits
The credit cost depends on the engine and resolution you select. These are the standard per-image rates:
Engine | 1K | 2K | 4K |
|---|---|---|---|
Fast | 1 credit | 2 credits | 3 credits |
Pro | 1 credit | 2 credits | 4 credits |
Max | 3 credits | 5 credits | Not available |
Your actual cost may depend on the settings available in your workspace. Always check the Run button before generating, as it shows the cost for your selected settings.
For more details, see the Caimera Credits Usage Guide.
Using model images as references
You can use a model image as a reference when generating or editing images in a workflow.
- Choose a model image: use a model image from Caimera's model roster or upload an image of your own model.
- Choose the right framing: a clear waist-up or close-up image of the model is recommended when you want the result to follow the model's appearance.
- Connect it as a reference: connect it to Create Image after the image you're editing (or as any image when generating from scratch).
- Describe the intended result: explain how you want the model to appear in the final image and identify the model reference clearly in your prompt.
If you need an image from Caimera's model roster, contact our team to request one for use in your workflow.
Note: custom model creation and training aren't supported in Workflows yet. You can still use an existing model image as a reference.
Common uses
- Swap the background: connect the product photo and prompt "Place $product on a plain white studio background. Keep the product exactly as it is."
- New angles: several Create Image nodes fed by the same input, each asking for one view, such as front, back, side or three-quarter.
- Style a garment on a model: connect the model image first and the garment second, then tag both in the prompt.
- Generate, refine, upscale: a Create Image node with only a prompt, then a second Create Image node that edits the result, then Upscale.
FAQs:
- What happened to Generate Image and Edit Image?
They're now one node, Create Image. Leave the image input empty to generate, or connect an image to edit it. Your existing workflows were switched over automatically with their prompts, settings and connections intact.
- Which connected image gets edited?
The first one (Image 1). Every image after it is used as a reference to guide the result.
- How many images can I connect?
Up to 6: the image to edit plus up to 5 references, or up to 6 references when generating from a prompt.
- Can I use my own model images?
Yes. Add a clear waist-up or close-up image of your model as a reference. You can use an image from Caimera's model roster by requesting it from our team, or upload one of your own. Custom model creation and training aren't supported in Workflows.
- Which engine should I choose?
Use Fast for quick tests and brainstorming, Pro for a balance of quality and speed, and Max for detailed textures, prints, and complex edits.
- Can Create Image process multiple images?
Yes. Connect it to a Batch Input node to apply the same prompt and settings to each image in the batch.
- Do AI assistants connected through MCP use this node?
Yes. Claude, ChatGPT and other assistants connected through MCP now use Create Image when they build workflows for you.
Related Articles:
📚 Workflows node reference: Every node you can add to a workflow.
📝 Upload & Edit Your Images on Caimera: The standalone editing tool.
🗂️ Create from Scratch: Generate patterns, sketches and graphics from text.
🖍️ Guide to Image Generation and Edit Models: How Caimera's models compare.
Have a question or running into an issue? Reach us via the chat bubble or at contact@caimera.ai
Updated on: 28/10/2026
Thank you!