Image Generator Component

Use the Image Generator component to create new images from text prompts (text-to-image) or modify existing ones (image-to-image). It leverages advanced models to bring your creative ideas to life.

Why this matters

Visual content is highly engaging. With the Image Generator, your agent can create custom graphics, illustrations, and art on the fly, opening up a world of possibilities for dynamic content creation, personalization, and enhanced user interaction.

What You’ll Configure

• Model and Generation Settings
• Define Inputs
• Handle the Generated Output
• Model Verification and Troubleshooting
• Best Practices
• Troubleshooting Tips
• What to Try Next

Step 1: Select a Model and Configure Settings

Choose a model and fine-tune its parameters to control the style, quality, and dimensions of the generated image.

Setting	Description
Model	Select from available models like GPT Image 1 (DALL-E 3), FLUX Dev, or FLUX Schnell.
Image Dimensions	Choose the aspect ratio and size of the output, such as 1024x1024 or 1792x1024.
Image Quality	Adjust the quality level of the generated image. Higher quality may take longer to generate.
Output Format	Select the desired format for the output image (e.g., PNG, JPG).
Strength	(FLUX models only) Controls how much the input image influences the output in an image-to-image task.
Negative Prompt	(FLUX models only) Specify concepts, objects, or styles to exclude from the generated image.

Step 2: Define Inputs

Dynamic Prompts

You can create additional text inputs for the component and reference them in the main Prompt field. For example, you could have an input named subject and write a prompt like: "An oil painting of a {{subject}} in a surrealist style."

Provide the creative direction for the image generation through prompts and optional attachments.

Input	Required?	Description
Prompt	Yes	The primary text description of the image you want to create.
Attachment	No	An optional image file (binary input) used for image-to-image generation.

Multiple Image Inputs

For models that support it, you can add multiple binary inputs. The component will automatically recognize and use them in the generation process without needing to reference them explicitly in the prompt.

Step 3: Handle the Generated Output

The component produces a single output containing the newly created image.

Output	Description	Data Structure
Output	A reference to the generated image file, ready to be used in subsequent steps.	Binary

Model Verification and Troubleshooting

Verifying Available Models

• Open the Model dropdown to confirm which models are currently available
• Expected options include GPT Image 1, FLUX Dev, and FLUX Schnell

Model-Specific Settings

• Each model provides different configuration options in the Advanced Settings panel
• FLUX models offer additional settings:
  • Strength — controls how closely output resembles the input image
  • Negative Prompt — exclude specific elements from output
• GPT Image 1 uses a different set of parameters and does not include Strength or Negative Prompt

Troubleshooting Model Selection

• If the model dropdown is empty or missing expected entries, check your internet connection or refresh the canvas
• Still not showing? Contact support for help diagnosing model availability

Best Practices

• Write Detailed Prompts: The more specific and descriptive your prompt, the better the result. Include details about the subject, style, colors, lighting, and composition.
• Iterate with Negative Prompts: Use the Negative Prompt (on FLUX models) to refine your images by removing unwanted elements. For example, if you're getting extra limbs, add "mutated, extra limbs" to the negative prompt.
• Leverage Image-to-Image: For variations on a theme or style, provide an input Attachment. Use the Strength parameter to control how closely the output resembles the input.
• Choose the Right Model: FLUX Schnell is faster and great for rapid prototyping, while FLUX Dev or GPT Image 1 may offer higher quality for final results.

Troubleshooting Tips

If your images aren't turning out right...

• Generic or Uninspired Results: Your prompt is likely too simple. Add more detail, adjectives, and artistic direction. Instead of "a dog," try "a photorealistic image of a golden retriever puppy playing in a sun-drenched meadow."
• Unwanted Elements Appear: Use the Negative Prompt setting on compatible models to explicitly forbid those elements.
• Image is Too Different from Input (Image-to-Image): Your Strength parameter may be set too high. Lower the value to make the output adhere more closely to the input image.
• Component Fails or Times Out: Image generation can be resource-intensive. Ensure your connection is stable and check the status of the model provider if issues persist.

What to Try Next

• Use a GenAI LLM Component to generate a highly detailed image prompt, and feed that prompt into the Image Generator.
• Create an Agent Skill that takes a user's text input, generates an image, and displays it back to them, creating an interactive "art bot".