Seedream 4.5 API for Image Generation
Access Seedream 4.5 through the independent GoEnhance Open API. Generate from a 1–2,000 character prompt or add up to 14 reference images, choose 2K or 4K output, and retrieve each asynchronous result by task ID.
Create an API Key
Build Text-to-Image and Reference Workflows
Text-to-Image Generation
Set model to seedream-4-5 and provide a required prompt between 1 and 2,000 characters. Choose a documented aspect ratio or keep auto.
Reference-Guided Generation
Add image_list with 1–14 public image URLs for image-to-image guidance. Each referenced file can be up to 10 MB.
2K or 4K Output
Select image_size as 2K or 4K. GoEnhance documents the same per-image price for both tiers; final dimensions depend on the chosen ratio.
Validate the documented request fields before submitting a generation task.
Seedream 4.5 API Capabilities
Unified image endpoint
POST JSON to <code>https://api.goenhance.ai/api/v1/images/generations</code>.
Exact model selection
Set the required <code>model</code> field to <code>seedream-4-5</code>.
Prompt validation
Provide a required prompt containing between 1 and 2,000 characters.
Twelve ratio options
Choose <code>auto</code>, square, portrait, landscape, ultrawide, or tall ratios documented by GoEnhance.
Two resolution tiers
Set <code>image_size</code> to <code>2K</code> or <code>4K</code>; it combines with ratio to determine dimensions.
Async result handling
Store <code>img_uuid</code>, then poll job details or process HTTPS callback events.
Simple per-image billing across both documented output tiers.
Seedream 4.5 API Pricing
1.85 tokens per image
GoEnhance currently lists each Seedream 4.5 generation at 1.85 tokens.
$0.037 per image
The documented USD-equivalent price is $0.037 for each generated image.
Same rate for 2K and 4K
Both documented image-size tiers are billed at the same per-image rate; check current terms before large runs.
Integrate the Seedream 4.5 API
1
1. Create and secure an API key
Create a GoEnhance API key and send it in the
Authorization header from server-side code. 2
2. Build the request
Set
model: seedream-4-5, add the prompt, and optionally choose ratio, image size, reference images, and a public HTTPS callback URL. 3
3. Submit and resolve the result
POST the JSON payload to
https://api.goenhance.ai/api/v1/images/generations and save the returned img_uuid. Then poll GET /api/v1/jobs/detail or process callback events until the task reaches success or failure, and store the returned image URL.Use prompt and reference inputs in production-oriented image pipelines.
What Developers Build with Seedream 4.5
Product campaign assets
Generate polished product-scene concepts from structured briefs in the ratio required by each channel.
Reference-guided variations
Use approved visual references to explore related compositions, materials, lighting, or art direction.
Editorial illustration pipelines
Turn article briefs into image tasks, then route generated outputs through review and asset management.
High-resolution concept art
Create 2K drafts or 4K outputs for environments, storyboards, and preproduction exploration.
Localized creative variants
Programmatically adapt prompt details for different regions while keeping task tracking consistent.
Automated image workflows
Connect generation to callbacks, moderation, storage, approval, and downstream publishing systems.
Answers about pricing, inputs, output sizes, delivery, and ByteDance API intent.