Visual AI Image Generator (Text-to-Image + Image-to-Image)

Generate images visually in Pro Forge. Iterate with text-to-image and image-to-image references, keep every step on your board, and download anytime.

Visual AI Image Generator (built for iteration)

AI image generation is rarely “one prompt and done.” It’s an iterative process—tweaking wording, adjusting style, and building on what worked. Pro Forge lets you generate images visually on a node-based canvas, so you can track your prompt journey and refine it step by step.

Overview: what this is in Pro Forge

Pro Forge is a visual prompt builder and IDE. For image creation, you build an image-generation flow on a board using nodes—so your prompt, reference images, and outputs stay connected.

With Pro Forge you can:

  • Generate images from a text prompt (text-to-image)
  • Generate images using one or more reference images (image-to-image)
  • Feed generated images back into the next generation step to iterate in one connected flow
  • Keep images stored on the board and download them any time

Who this is for

  • Designers and creators iterating on style and composition
  • Marketers producing visuals and exploring concepts quickly
  • Product teams prototyping image ideas without losing version context
  • Anyone who wants a visual, repeatable workflow instead of prompt chaos

Key benefits

  • See the journey, not just the result: every prompt and image stays on the canvas as a connected flow.
  • Iterate faster: reuse prior outputs as references instead of starting over.
  • Less guesswork: keep your prompt text, reference inputs, and generated outputs together.
  • Built-in organization: boards act as “workspaces” where image experiments remain saved and revisitable.
  • Type-safe connections: image connections only connect to image inputs, and text connections to text inputs—reducing wiring mistakes.

How it works in Pro Forge

  1. Create a new Board
  2. Click the Add button
  3. Select Templates → choose “Image Generator”
  4. Enter your prompt (or connect prompt text)
  5. (Optional) Add reference images for image-to-image
  6. Generate and iterate
  7. Download images any time

Create a new Board Start from a project and open a new board (your visual canvas).

Click the Add button At the top center, click the circular “+” icon.

Select Templates → choose “Image Generator” This scaffolds a starting flow so you can go from a prompt to an image immediately.

Enter your prompt (or connect prompt text) Use a text prompt as your starting point (text-to-image).

(Optional) Add reference images for image-to-image Connect one or more images as references to guide style/content.

Generate and iterate Pro Forge creates an output image node, and you can feed the generated image into the next step to refine results in the same visual chain.

Download images any time Your generated images are stored on the board so you can revisit and download them later.

Try it out (demo)

Demo link is coming soon.

Can Pro Forge do both text-to-image and image-to-image?

Yes. The Image Generation node supports a text prompt input and also supports adding one or more image reference inputs so you can generate from text alone or guide generation using existing images.

Can I reuse a generated image to refine the next one?

Yes. Generated images appear as Image nodes on the board, and you can connect them as reference inputs for the next generation step—so iteration becomes a single connected workflow.

Where do my generated images live?

Images are stored in your board, so you can come back later and download them again (not just a one-time result).

What image sizes are available?

The Image Generation node includes size options such as 2K (default), 1080p, Square, Portrait, Landscape, 4:3, 3:4, and Mobile.

What model does Pro Forge use for image generation?

The Image Generation node uses an image model (default shown in-app as seedream-4-5-251128) and runs generation through Pro Forge’s backend.

What happens after I click generate?

Pro Forge automatically creates a new Image node for the output and the image loads in when ready. (The app polls for completion for up to about a minute.)

Can I connect anything to anything?

Connections are validated by type: text connects to text, and image connects to image. This helps keep your generation graph clean and prevents invalid wiring.

Is this meant for beginners or casual one-off prompts?

Pro Forge is built for people who want to iterate and maintain complex prompt workflows. If you only run one-off prompts occasionally, you may not need a visual canvas.