Text to image
Describe the subject, composition, lighting and style. Make a fresh product scene, portrait, illustration or social graphic without uploading a starting image.
Start from a promptByteDance Seedream · Available on XImagineAI
Turn an idea into a product photo, a poster or a transparent design asset. Create with Seedream AI from a text prompt, refine an image with references, or separate a finished composition into reusable layers—all in your browser.
Examples
Configure parameters and click Generate
Sign in to generate. Choose your settings and review the credit cost before submitting.
One idea, three ways to create
Describe the subject, composition, lighting and style. Make a fresh product scene, portrait, illustration or social graphic without uploading a starting image.
Start from a promptUpload 1–10 reference images and explain what to change and what to preserve. Try a new background, adjust a product scene or develop a consistent visual direction.
Edit with referencesUpload one image and separate it into a base image and individual PNG layers. Choose an output limit from 2 to 100, including the base image, and enable text recognition when needed.
Separate an imageFrom brief to image
Use Text to Image for a new idea. For an existing image, open Image to Image and select Seedream 5.5 Edit or Seedream 5.5 Layer Decomposition.
Write a specific prompt or add your reference images. Select the aspect ratio and resolution. For a cutout, choose PNG and a transparent background.
Review the credit estimate, generate and download your result. To make another edit, use the result as a new reference and describe the next change.
A useful starting point
Start with a clear subject and one purpose. Add framing, materials and light, then put any exact lettering in quotation marks. Review generated text before publishing.
01
Give a product a clear focal point and leave space for your campaign copy.
A studio product photograph of an amber glass skincare bottle on pale limestone, soft window light from the left, realistic glass reflections, warm ivory background, bottle on the right with empty space on the left for copy, no lettering or watermark.
For your own product, switch to Image to Image and upload its photo.
02
Keep the headline short and describe the hierarchy of the layout.
Design a vertical botanical workshop poster on warm cream paper. Large dark green headline at the top: "GROW TOGETHER". Smaller text below: "SATURDAY 10 AM". Fine botanical illustrations around the edges, generous margins, clear typography and an uncluttered center.
Choose 3:4 for a portrait poster and check every letter in the result.
03
Create an isolated element to place over a website, presentation or packaging design.
A single hand-painted lemon branch with two ripe lemons and delicate green leaves, watercolor texture, clearly separated leaf edges, isolated subject, no ground plane, no lettering, transparent background.
Choose PNG and set Background to Transparent before generating.
Make it fit your project
These are the Seedream 5.5 resolution options on XImagineAI. Use a lower tier to explore a concept, then choose 2K when you need more detail. This page does not offer a 4K setting.
Choose square, portrait, landscape or ultrawide formats, including 1:1, 3:4, 16:9 and 21:9. Image editing also supports Auto for the aspect ratio.
For transparent assets, select PNG and Transparent together; JPEG cannot preserve transparency. In layer decomposition, the format setting applies to the base image and the separated layers are PNG.
Know the cost before you create
The examples below use the current XImagineAI model settings. Your cost changes with resolution, reference count and, for layer decomposition, the output limit.
The generator shows the estimate for your selected settings before submission. Layer decomposition displays an upper estimate based on the output limit. Review the cost shown in the generator before starting the task.
Can I try Seedream 5.5 free? If your account has promotional credits, you can use them when your balance covers the displayed cost. Availability and amounts depend on current account offers; generation is not unlimited.
View credit packs and plansSeedream 5.5 is ByteDance's image generation and editing model. On XImagineAI, you can use it for text-to-image creation, reference-based image editing, transparent PNG assets and layer decomposition.
Use the generator on this page with an XImagineAI account and enough credits for your chosen settings. There is no app to install and you do not need your own API key. XImagineAI is an independent platform, not ByteDance's official website or Dreamina.
Yes. Choose Image to Image and Seedream 5.5 Edit, upload up to 10 references and describe the edit. State what should remain unchanged, such as the product shape, face, camera angle or background. Review each result before continuing.
Use the text-to-image or image-editing model, choose PNG as the output format and set Background to Transparent. Describe the isolated subject in your prompt. A prompt alone does not replace the background setting.
The Seedream 5.5 options on XImagineAI are 1K, 1.5K and 2K. Layer decomposition also has an Auto size option. Check the settings for the selected model rather than assuming another Seedream version's limits apply.
This page provides a browser-based generator. You do not need API credentials or a ComfyUI workflow to use it. For a programmatic integration, consult your chosen API provider's documentation and pricing separately from XImagineAI credits.
No. Seedream is for images; Seedance is a video model family. If you want to animate a still image, use an image-to-video tool after creating your image.
Use the same brief and, for editing, the same references. Compare the accuracy of the requested change, text legibility, subject consistency and credit cost. Model names and features vary by version; select Seedream 5.5 explicitly for the controls described on this page.
Explore Seedream and other image models in one workspace.
Try another model with your image brief.
Explore video creation for your next project.
Find the image or video model for your task.
Start with a prompt, a reference or an image you want to take apart.
Create with Seedream 5.5