The image workspace opens with generation controls on the left and the output panel on the right. All Seedream 5.0 Pro settings, including model selection, prompt, aspect ratio, resolution, and output format, are configured from the left panel before you run a generation.

Click the model dropdown at the top of the panel. Seedream 5.0 Pro appears as two separate entries, one for Text to Image and one for Image to Image. Select the Text to Image entry to generate images from a written prompt.

The type dropdown lets you switch between Text to Image and Image to Image within the same session without changing the selected model. Confirm that Text to Image is selected before writing your prompt.

Enter your prompt in the text field. Seedream 5.0 Pro handles complex scene descriptions accurately, including lighting conditions, material details, and typographic elements rendered inside the image. Describe subject, setting, lighting, and style for the most consistent results.

The input images section lets you upload reference photos to guide visual style or subject consistency. Use this to anchor the generated image to a specific subject, costume, or setting visible in the reference. This step is optional when generating from a text prompt alone.

Seedream 5.0 Pro supports 14 aspect ratios: 1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, and 21:9. Choose 16:9 for widescreen composites, 9:16 for vertical formats, 4:3 for classic photography proportions, or 1:1 for square outputs.

Select 1k (45 credits) for composition checks and iterative drafts. Select 2k (90 credits) when the image is going into final production, print, or a high resolution display context.

Choose JPEG for smaller file sizes suited to web delivery. Choose PNG when you need lossless quality or when the image feeds into a pipeline requiring a transparent capable format, such as a background removal step.

The workspace shows the credit cost for your selected resolution tier before the generation runs. Credits are deducted at a fixed rate per tier, with no variable charges beyond what is displayed. Confirm and click Generate.

The generated image appears in the output panel on the right. From here you can download it directly, run another variation with the same settings, or adjust the prompt and regenerate.

Use Save as Actor to store the image as a reusable visual reference for anchoring subject identity or appearance in future generations. Use Save as Character to register it as a named character for identity consistent workflows across Text to Image, Image to Image, and video generation nodes.

Add a Prompt node to the canvas and connect its Mauve output port to the input of a Text to Image node configured with Seedream 5.0 Pro. Connect the Lavender image output port to a Result or Image Viewer node. Press Generate to run the pipeline and collect the output.

The Result node displays the generated image on the canvas. Extend the pipeline by connecting the image output to a background removal node, an image enhancer, or an image-to-video node. For a worked example combining image generation with background removal, see the BEN2 Background Remover tutorial.

From the image workspace, click the model dropdown and choose the Image to Image entry for Seedream 5.0 Pro. The interface updates to show an image upload section below the prompt field.

The model list shows the Image to Image variant of Seedream 5.0 Pro as a separate entry from the Text to Image entry. Select it here to enable the reference image upload controls.

In the type dropdown, confirm Image to Image is selected. You can switch back to Text to Image at any point in the same session without losing your uploaded reference images or prompt text.

Upload your reference images in the area below the prompt field. Seedream 5.0 Pro accepts up to 10 reference photos in a single request. Write a prompt describing the transformation you want. A prompt such as "change the lighting to golden hour, warm tones, preserve the subject and composition" gives the model clear direction while keeping the original structure intact.

The first reference image determines the default output aspect ratio if no aspect ratio is set manually. Each reference image beyond the first adds 3 credits to the base cost of the selected resolution tier.

Select 1k (45 credits base) for fast iterative editing or 2k (90 credits base) for final output at full resolution. The resolution tier applies to the generated image, independent of the size of your uploaded reference photos.

The safety checker filters generated output against platform content policies. It is active by default for all Image to Image generations and the toggle is visible in the generation controls panel before you run the request.

Click Generate. After the image is produced, use Save as Actor to store the result as a visual reference for future generations, or use Save as Character to register it as a named character for consistent character workflows across the studio.

In the Nodes Graph Editor, add a Prompt node and one or more Image Upload nodes to the canvas. Connect the Prompt node's Mauve output and each Image Upload node's Lavender output to the Image to Image node configured with Seedream 5.0 Pro. Connect the image output to a Result or Image Viewer node and press Generate.

Up to 10 Image Upload nodes can be connected as reference inputs to the same generation node. For pipelines that combine image generation with video output, see the FLUX 2 tutorial for additional node pipeline patterns.