Seedream 4.5Image API

Seedream 4.5 APIfor developers.

Generate polished commercial visuals, editing-consistent updates, and richer multi-image compositions through ByteDance's newest Seedream image model on ImaRouter.Built for teams that care about subject preservation, lighting consistency, portrait refinement, and cleaner small-text rendering.

Creative Direction

Visual references for image-led video workflows

Selected frames and stills that help explain how source images, reference material, and finished outputs can shape a Happy Horse video workflow.

Editorial beauty product image used as Seedream 4.5 visual direction.

Beauty Still

Editorial beauty product visual

A premium editorial-style still that fits Seedream's strength in polished, commercially presentable image outputs with clean material rendering.

Jewelry product display reference image for Seedream 4.5 composition workflows.

Composition Reference

Jewelry product display setup

A composition-led product setup that matches Seedream's stronger multi-image and layout-aware visual generation workflows.

Generation quality

Commercial-grade stills

Built for higher-fidelity image generation with stronger aesthetics and cleaner material rendering

Edit consistency

Preserve details better

More reliable preservation of subject details, color tone, and scene lighting during edit-like workflows

Text behavior

Sharper small text

Improved rendering for labels, packaging copy, UI-like layouts, and denser commercial text elements

Portrait quality

Refined faces and skin

Better portrait refinement for creator, beauty, and fashion-facing image use cases

Composition

Stronger multi-image blends

More capable when several references need to resolve into one coherent output

Pricing

$0.04 / output image

Reference pricing from the current Seedream 4.5 model description, regardless of final image size

Available Endpoints

Start building with the Seedream 4.5 API

Multiple endpoints for text-to-video, image-to-video, fast preview flows, and async job retrieval. This section is laid out more like a product catalog than raw docs so users can scan what to use first.

Seedream 4.5 beauty product reference output.
NewGenerate

Endpoint

Generate

/v1/image/seedream-4.5/generate

Core endpointText-to-imageCommercial visualsQuality-first

Generate a new image from a prompt with commercial-style quality, richer detail retention, and stronger visual polish.

Best for: Best for brand visuals, beauty product stills, packaging concepts, premium ad creative, and reference-light concept generation.

Seedream 4.5 transformation style edit reference.
NewEdit

Endpoint

Edit

/v1/image/seedream-4.5/edit

Editing consistencyPreserve detailsPortrait refinementCommercial edit

Refine an existing image while preserving more subject detail, lighting continuity, and color tone consistency.

Best for: Useful for scene cleanup, product revisions, packaging updates, portrait polishing, and preserve-and-change commercial workflows.

Seedream 4.5 multi-image composition reference.
NewCompose

Endpoint

Compose

/v1/image/seedream-4.5/compose

Multi-imageCompositionReference blendContext-aware

Combine multiple reference images into one output with stronger composition logic, consistency, and visual coherence.

Best for: Best for campaign compositions, style-reference blending, product placement, or image systems built from several source inputs.

New

Endpoint

Job Status

/v1/jobs/{jobId}

Async statusPollingProduction flow

Track whether a generation, edit, or composition job is queued, running, completed, or failed in ImaRouter's async image flow.

Best for: Needed for production applications that submit asynchronous image jobs and then retrieve final output after completion.

Get started today

Ready to integrate Seedream 4.5?

Try the API directly in the console, or reach out to the team for onboarding, pricing, and enterprise setup.

API Documentation

How to get access to Seedream 4.5 API

Seedream 4.5 fits the same async image workflow pattern as the rest of the ImaRouter image stack: submit the request, keep the job id, then fetch the result once the image is complete.

Selected endpoint

/v1/image/seedream-4.5/generate

Start with prompt-based generation, then move to edit or compose mode when the workflow depends on stronger consistency, portrait refinement, or multi-reference composition.

Best for brand visuals, beauty product stills, packaging concepts, premium ad creative, and reference-light concept generation.

const apiKey = process.env.IMAROUTER_API_KEY;

async function createSeedreamVisual() {
  const createResponse = await fetch("https://api.imarouter.com/v1/image/seedream-4.5/generate", {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${apiKey}`,
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      prompt: "A premium skincare bottle on glossy black stone, soft violet rim light, delicate mist, crisp packaging typography, luxury campaign composition, and polished commercial reflections",
      imageSize: "portrait_4_3",
      quality: "high",
      numImages: 1,
      outputFormat: "png"
    })
  });

  const job = await createResponse.json();

  let status = "queued";
  while (status !== "completed") {
    await new Promise((resolve) => setTimeout(resolve, 3000));

    const statusResponse = await fetch(`https://api.imarouter.com/v1/jobs/${job.id}`, {
      headers: {
        "Authorization": `Bearer ${apiKey}`
      }
    });

    const jobState = await statusResponse.json();
    status = jobState.status;

    if (status === "failed") {
      throw new Error(jobState.error ?? "Seedream 4.5 generation failed");
    }

    if (status === "completed") {
      return jobState.output[0].url;
    }
  }
}

Async flow

  1. 1

    Choose whether the request is a new generation, a consistency-sensitive edit, or a multi-image composition.

  2. 2

    Submit the prompt together with any source image URLs or reference images and keep the returned job id.

  3. 3

    Poll the job endpoint until the image reaches completed or failed state.

  4. 4

    Read the final output URL and push the approved visual into your own campaign, commerce, or publishing workflow.

What Makes It Different

What makes the Seedream 4.5 API different

This section is laid out to read more like a product narrative than a feature list. Each row shows a capability, why it matters, and what that looks like in a real workflow.

Preview

Stronger editing consistency

That makes it more useful for teams that need revisions and transformations without losing the fidelity of the original asset.

Capability

Stronger editing consistency

Seedream 4.5 is positioned as a major step up over Seedream 4.0 in editing consistency, especially when preserving subject details, lighting direction, and scene tone.

That makes it more useful for teams that need revisions and transformations without losing the fidelity of the original asset.

Example scenario

A campaign team updates packaging, props, or styling while keeping the product silhouette, material finish, and scene lighting intact.

Capability

Portrait refinement

The model improves portrait refinement, making it more suitable for beauty, creator, and fashion-oriented imagery where face quality matters under review.

Teams can use it for skincare campaigns, creator stills, social ads, and premium portrait-led product visuals.

Example scenario

A beauty brand rerenders a hero still with cleaner skin detail, more believable eye light, and stronger premium polish.

Preview

Portrait refinement

Teams can use it for skincare campaigns, creator stills, social ads, and premium portrait-led product visuals.

Preview

Sharper small-text rendering

This is especially useful for product labels, beauty packaging, promo stills, and educational or commerce-facing layouts.

Capability

Sharper small-text rendering

Seedream 4.5 improves small-text rendering, which is critical when packaging copy, interface-like labels, ingredient callouts, or campaign text need to survive generation.

This is especially useful for product labels, beauty packaging, promo stills, and educational or commerce-facing layouts.

Example scenario

A brand team generates packaging mockups and wants ingredient labels or product names to remain readable without rebuilding the layout manually.

Capability

More capable multi-image composition

Multi-image composition is described as significantly strengthened, making Seedream 4.5 more practical for reference-heavy visual workflows.

You can combine style references, product shots, scene context, and subject imagery into one more coherent output.

Example scenario

A commerce workflow merges a clean packshot, a background scene, and a lighting reference into a polished catalog-ready ad visual.

Preview

More capable multi-image composition

You can combine style references, product shots, scene context, and subject imagery into one more coherent output.

Unified API Platform

Two API tiers for different use cases

Pick the right balance of quality, speed, and cost for your workflow. The section stays data-driven, but the presentation is closer to a clean product comparison table.

Feature
Standard generationRecommended
Consistency-sensitive edit
Reference composition
Best forPrompt-led commercial image creationPreserve-and-change workflowsMulti-image commercial workflows
SpeedBalancedModerateModerate
QualityStrong aesthetics and polished visualsOptimized for stronger detail retentionBest when several references shape one output
Cost$0.04 / image$0.04 / image equivalent$0.04 / image equivalent
Recommended useBrand visuals, campaign stills, product shots, and polished first-pass concept creationRefining portraits, preserving product details, relighting, and packaging-safe revisionsProduct placement, style blending, scene composition, and campaign systems that start from several inputs
API endpoints/v1/image/seedream-4.5/generate/v1/image/seedream-4.5/edit/v1/image/seedream-4.5/compose

Use Cases

Industries using the Seedream 4.5 API

This section keeps the same reusable data model, but the presentation is closer to a grid of industry cards instead of long narrative boxes.

Brand teams, agencies, and creative ops

Beauty and skincare campaign stills

Generate polished portrait-led product imagery with stronger facial refinement, premium lighting, and cleaner packaging presentation.

Seedream 4.5 is a stronger fit when portrait quality and premium commercial polish matter in the same frame.

Merchandising teams and product marketers

Packaging and label visualization

Prototype bottles, cosmetics, product cards, and packaging visuals where small copy and clean material rendering both matter.

Improved small-text rendering makes it more practical for label-heavy commercial assets than generic image models.

Creative ops and design systems teams

Consistency-focused image revisions

Revise approved campaign assets while preserving core subject details, scene lighting, and color tone rather than rebuilding from scratch.

Editing consistency is one of the model's clearest improvements and maps directly to production revision workflows.

Commerce teams and visual production pipelines

Reference-heavy commercial composition

Blend product images, scene references, and style direction into one controlled output for ads, PDP stills, and premium commerce visuals.

Stronger multi-image composition makes it easier to operationalize reference-led visual creation instead of depending on one-prompt generation alone.

Performance marketers and creator tools

Creator and portrait-led ad generation

Create portrait-forward ad stills, creator campaign visuals, and social-first image outputs with cleaner refinement and scene consistency.

Portrait refinement matters most in ad surfaces where the face is the conversion anchor and small rendering flaws break trust quickly.

Agencies and in-house design teams

Iterative commercial art direction

Use one model path for initial visual direction, then push the same asset through edits and compositions as the campaign sharpens.

The model is useful when the work is not only generation, but also revision, refinement, and compositional iteration over time.

Examples

Seedream 4.5 API examples

This section supports future video or image previews, but it also renders cleanly from prompt-only data. That keeps the template reusable even when a page launches before media assets are ready.

Preview

Media coming soon

This example will be updated with a visual preview.

Premium skincare hero still

Commercial product generation

A prompt like this is useful when you need clean material rendering, luxury atmosphere, and small packaging text that stays believable.

A premium skincare bottle on glossy black stone, soft violet rim light, realistic condensation, clean serif packaging text, subtle mist, and a luxury editorial campaign finish.

beautyproductcommercial
Preview

Media coming soon

This example will be updated with a visual preview.

Portrait refinement edit

Consistency-led edit

This is the kind of edit where Seedream 4.5's consistency improvements matter more than raw generation speed.

Preserve the model pose and styling, refine the skin texture naturally, improve eye detail, keep the lighting direction intact, and make the cosmetic packaging text cleaner without changing the composition.

portraiteditbeauty
Preview

Media coming soon

This example will be updated with a visual preview.

Jewelry product composition

Multi-image composition

A reference-heavy visual workflow where composition quality and coherence matter more than one-shot prompt generation alone.

Blend the jewelry product from image 2 into the pedestal scene from image 1, keep the reflections premium, preserve stone texture realism, and maintain a clean luxury campaign composition.

jewelrycompositionreference-led
Preview

Media coming soon

This example will be updated with a visual preview.

Localized packaging still

Small-text rendering

A useful pattern when packaging copy and readability are part of the asset instead of an afterthought.

Generate a beauty packaging still for a Korean skincare serum with readable product label text, ingredient callouts, and clean studio lighting while keeping the scene premium and minimal.

packagingtext renderingcommerce

How To Use This API

How to use Seedream 4.5 API

This quick-start walkthrough is written to rank for integration-style searches while staying concise enough for busy developers and operators.

  1. 1

    Create your ImaRouter account

    Set up your ImaRouter account so you can access the playground, dashboard, and API credentials from one place.

  2. 2

    Choose generate, edit, or compose mode

    Use generate for net-new visuals, edit when detail preservation matters, and compose when several references need to resolve into one output.

  3. 3

    Write consistency-aware prompts

    Be explicit about what must remain unchanged, especially when preserving subject detail, lighting tone, label readability, or commercial polish.

  4. 4

    Submit the request and keep the job id

    Store the returned job id and treat the image workflow as asynchronous, especially when integrating into production campaign or commerce systems.

  5. 5

    Review output and rerun selectively

    Use the first result to validate composition, text, portrait quality, and consistency, then rerun targeted variations instead of restarting from a vague prompt.

FAQ

Frequently asked questions about Seedream 4.5 API

FAQs stay compact and skimmable here. The content is still data-driven for SEO, but the layout is cleaner and less visually heavy.

What is Seedream 4.5 API?

It is ImaRouter's programmable interface for ByteDance's Seedream 4.5 image model, designed for high-fidelity image generation, stronger editing consistency, portrait refinement, small-text rendering, and richer multi-image composition.

What does Seedream 4.5 improve over Seedream 4.0?

The model description highlights stronger editing consistency, better preservation of subject detail, lighting and color tone, sharper portrait refinement, better small-text rendering, and significantly improved multi-image composition.

Is Seedream 4.5 better for generation or editing?

It is useful for both, but one of its clearest differentiators is that consistency-sensitive editing and reference-driven workflows are stronger than in prior versions.

What kinds of teams should use Seedream 4.5?

It fits beauty, fashion, packaging, premium commerce, and creative ops teams that need quality-first image output with stronger subject preservation and better commercial polish.

Does Seedream 4.5 handle small text better?

Yes. The referenced model description explicitly calls out improved small-text rendering, which makes the model more practical for labels, packaging, and commercial stills with dense visual copy.

How much does Seedream 4.5 cost?

The referenced public description lists pricing at $0.04 per output image regardless of size.

Why use ImaRouter for Seedream 4.5 instead of wiring provider-specific image routes?

It gives you one production-safe integration pattern for generation, edits, and compositions while keeping async job handling, API credentials, and future routed expansion under one platform.

Can Seedream 4.5 support multi-image commercial workflows?

Yes. Multi-image composition is described as significantly strengthened, which is one of the model's clearest advantages for reference-heavy commercial generation pipelines.

Model Directory

Browse the full model market before you choose your route.

Use the `/models` catalog to scan providers, modalities, reasoning support, context windows, and pricing metadata from a local OpenRouter snapshot. It is the fastest way to compare what exists before you decide which models should be prioritized on ImaRouter.

Get Started

Validate Seedream 4.5 on ImaRouter, then move the same image workflow into production

Use the playground to validate premium generation, consistency-led edits, and reference compositions, then move the same pattern into production with ImaRouter image endpoints and async jobs. Use one API surface for 80+ models across five modalities, with transparent routing, automatic failover, and fast new-model onboarding.