# Seedream 5 Pro - Core.Today AI API

> ByteDance's flagship Seedream 5.0 Pro image generation model, with built-in reasoning, precise instruction following, and multi-reference support for up to 10 images. Offers 1K and 2K resolution output with higher fidelity than Seedream 5 Lite.

- **Provider**: ByteDance
- **Model ID**: bytedance/seedream-5-pro
- **Category**: Image Generation
- **Credits**: 210 per image, tiered by resolution
- **Speed**: Medium
- **Quality**: Ultra

## Features
- Built-in reasoning for precise instruction following
- Multi-reference generation with up to 10 input images
- 1K and 2K resolution output
- Aspect ratio auto-match to the input image
- Higher fidelity than the Seedream 5 Lite sibling model
- PNG or JPEG output format

## Use Cases
- Production-grade marketing visuals requiring precise instruction following
- Multi-reference product photography combining up to 10 source images
- High-resolution (2K) hero images for campaigns
- Image-to-image editing guided by multiple reference photos
- Batch content production that must preserve the input image's aspect ratio

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - Text prompt for image generation. Maximum 4000 characters. BytePlus recommends keeping prompts under 600 English words for best results.

### Optional
- **image_input**: array (default: []) - Input image(s) for image-to-image generation. List of 1-10 reference images for single or multi-reference generation.
- **size**: string (default: 2K) - Image resolution: 1K (~2 megapixels) or 2K (~4 megapixels).
  Options: 1K, 2K
- **aspect_ratio**: string (default: match_input_image) - Image aspect ratio. Use 'match_input_image' to automatically match the input image's aspect ratio.
  Options: match_input_image, 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9
- **output_format**: string (default: png) - Output image format.
  Options: png, jpeg

## Examples

### Multi-Reference Product Composite
Combine multiple product reference photos into a single cohesive studio shot
```json
{
  "model": "bytedance/seedream-5-pro",
  "input": {
    "prompt": "Combine these three sneaker reference photos into a single studio product shot on a seamless white background, consistent lighting from the upper left, and a soft shadow beneath each shoe",
    "image_input": [
      "https://example.com/sneaker-1.jpg",
      "https://example.com/sneaker-2.jpg",
      "https://example.com/sneaker-3.jpg"
    ],
    "size": "2K",
    "aspect_ratio": "1:1"
  }
}
```

### High-Resolution Marketing Banner
Text-only prompt generating a wide banner at 2K resolution
```json
{
  "model": "bytedance/seedream-5-pro",
  "input": {
    "prompt": "A minimalist tech product launch banner featuring sleek wireless earbuds floating above a reflective dark surface, dramatic rim lighting, deep navy background with a subtle gradient",
    "size": "2K",
    "aspect_ratio": "21:9",
    "output_format": "png"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Set size to 1K when maximum resolution isn't required — it's the lighter tier below 2K
- Provide up to 10 reference images in image_input for consistent multi-subject composites instead of relying on prompt text alone
- Use aspect_ratio: match_input_image when editing a reference photo to preserve its original framing
- Keep prompts under roughly 600 English words per BytePlus guidance — longer prompts don't improve fidelity

## Documentation
https://replicate.com/bytedance/seedream-5-pro