Official introduction
BytePlus Lumina has a Seedream 5.5 introduction with details on image creation and editing. Follow the official page for product and launch updates.
Seedream 5.5 brings local editing, realistic compositions, multilingual layouts and transparent assets into one image creation workflow. Refine a product image, adapt a campaign or prepare elements for a layered design. API access through EvoLink is coming soon; subscribe for a launch update or explore the Seedream models you can use today.
BytePlus Lumina has a Seedream 5.5 introduction with details on image creation and editing. Follow the official page for product and launch updates.
The Seedream 5.5 API details will cover the model ID, input and output options, limits and pricing.
Seedream 5.5 is coming soon to EvoLink. Subscribe to receive an update when access opens.
When EvoLink access opens, this page will include integration details, supported options and pricing.
Get notified when Seedream 5.5 API access opens on EvoLink.
We will only email you about Seedream 5.5 availability.
Follow the discussion Join EvoLink on Discord
Find out how to call Seedream 5.5 through EvoLink, which image options are supported and what requests cost. You can start current projects with an existing Seedream model while you wait.
Clear instructions and well-chosen references make it easier to develop an image without losing the details that matter.
Identify the object, its current state, the target state and the allowed edit region. Make every instruction self-contained and repeat the details that must remain unchanged.
Specify which image supplies the subject, product details, pose, style or layout. For consecutive edits, pass the previous output as the next input.
Quote all required text, numbers and punctuation. Define line breaks, alignment and hierarchy. Separately describe transparent background regions, semi-transparent fills and transparent materials.
When choosing a model, consider output settings, request costs, waiting time, retries and manual cleanup together. Base your budget on the assets you can use.
Develop product visuals, localized campaigns and reusable design elements from the same creative direction.
Create product and campaign variants by changing the background, camera angle or model styling around the same product. Give the product’s shape, material, logo and colors as fixed references, with a clear perspective, scale and lighting direction.
Build posters, labels and information graphics with a clear title, main visual, information area and brand placement. Provide the exact headline, body copy and numerals, then adapt the same layout to different languages.
Adjust expressions, accessories or a selected portrait area, or develop a sequence of storyboard frames. Define the reading order, framing, camera angle and lighting while keeping the character’s identity and appearance consistent.
Create stickers, icons and decorative elements for different backgrounds, or separate a composition for further changes. Reuse individual elements, replace text and adapt a design without rebuilding every part.
Local edits, realistic images, multilingual text, transparent assets and layered revisions support the steps from a first image to the finished design. Supported API options will be listed when EvoLink access opens.
Change a selected object or region while keeping the rest of the composition in place. Refine accessories, expressions or product colors across successive revisions.
Create portraits and product scenes with natural expressions, proportions, materials and lighting. Combine reference images with attention to scale, perspective and light direction.
Arrange headlines, small print and multilingual copy in posters, labels and information graphics, with a clear visual hierarchy and reading order.
Create reusable assets with transparent backgrounds and semi-transparent areas for stickers, product cutouts and decorative overlays.
Separate a composition into background, visual and text elements for further changes, text replacement and language variants.
Explore the Seedream models already available through EvoLink’s unified API gateway. Choose one for your current image generation or editing work, then consider Seedream 5.5 when its access and supported options are available.
Match the model to your product images, edits or layouts. Keep your current model when it already delivers the results you need.
Include request costs, retries and manual retouching in your budget for the finished image.
Keep model selection configurable and check the selected version’s inputs and output settings before changing your integration.

Seedream 5.0 Pro
Image generation and editing, with input options and pricing on the model page.
View model

Seedream 5.5 is coming soon to EvoLink. Subscribe to the launch alert to hear when API access opens.
Pricing will be available on this model page when Seedream 5.5 launches on EvoLink.
The Seedream 5.5 API documentation will list its model ID, input formats, reference-image options, output settings and request limits when EvoLink access opens.
Use each round’s output as the next input. Name the object to edit, its current state, the desired change and the details to preserve. Save intermediate versions so you can return to a preferred result.
Specify the transparent background, visible elements and any semi-transparent or cutout areas. Describe the appearance of materials such as glass separately. Check the exported alpha channel and preview the asset on different backgrounds.
Quote every required string exactly, including punctuation and numbers. Specify text hierarchy, location, alignment and line breaks, then check both character accuracy and layout in each language. For text replacement, define the old-to-new text mapping and preserve unrelated design elements.