> ## Documentation Index
> Fetch the complete documentation index at: https://docs.anyfast.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Seedream 5.0 Lite i2i

> Edit images and create coherent image sets from references with ByteDance Seedream 5.0 Lite.

Seedream 5.0 Lite i2i is ByteDance's reference-image editing workflow for single-image and coherent multi-image output. It combines precise editing, multiple references, visual instructions, and consistent image-set generation.

Use model ID `seedream-5-0-lite-260128`.

<Info>
  For specific content requests, select the **Special-Ns** resource group.
</Info>

## Key capabilities

* **Image editing** — Add, remove, replace, or restyle content while preserving the requested parts of the source image.
* **Multi-image reference** — Combine subjects, products, styles, or layouts from multiple reference images.
* **Batch image output** — Generate coherent storyboards, comics, brand assets, or other related image sets.
* **Visual instruction editing** — Use arrows, boxes, masks, or doodles in the input image to identify the area to edit.
* **Knowledge-rich visuals** — Create diagrams, formulas, educational illustrations, and dense information graphics.

## Image editing and multi-image reference

Pass `image` as one URL/Base64 data URI or an array of references. Clearly identify each image in the prompt when using multiple inputs.

```bash cURL theme={null}
curl https://www.anyfast.ai/v1/images/generations \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedream-5-0-lite-260128",
    "prompt": "Replace the clothing in image 1 with the outfit from image 2 while preserving the person, pose, lighting, and background from image 1",
    "image": [
      "https://example.com/person.png",
      "https://example.com/outfit.png"
    ],
    "size": "2K",
    "output_format": "png",
    "watermark": false
  }'
```

## Batch image output

Set `sequential_image_generation` to `auto` and use `max_images` to cap the number of images.

```json theme={null}
{
  "model": "seedream-5-0-lite-260128",
  "prompt": "Generate four coherent cinematic storyboard frames showing an astronaut repairing a spacecraft, encountering a meteor shower, escaping the impact, and returning safely",
  "size": "2K",
  "sequential_image_generation": "auto",
  "sequential_image_generation_options": {
    "max_images": 4
  },
  "output_format": "png",
  "response_format": "url",
  "watermark": false
}
```

## Parameters

| Parameter                                        | Type            | Required | Description                                                                                         |
| ------------------------------------------------ | --------------- | -------- | --------------------------------------------------------------------------------------------------- |
| `model`                                          | string          | Yes      | Must be `seedream-5-0-lite-260128`.                                                                 |
| `prompt`                                         | string          | Yes      | Text instruction for generation or editing.                                                         |
| `image`                                          | string or array | Yes      | Reference image URL/Base64 data URI, or an array of references. Supports up to 14 reference images. |
| `size`                                           | string          | No       | Resolution tier `2K`, `3K`, or `4K`, or a supported explicit pixel size.                            |
| `seed`                                           | integer         | No       | Random seed used to make repeated requests more consistent.                                         |
| `sequential_image_generation`                    | string          | No       | `auto` enables batch image output; `disabled` returns one image.                                    |
| `sequential_image_generation_options.max_images` | integer         | No       | Maximum images to return when batch output is enabled.                                              |
| `response_format`                                | string          | No       | `url` or `b64_json`. Generated URLs expire after 24 hours.                                          |
| `output_format`                                  | string          | No       | `png` or `jpeg`.                                                                                    |
| `watermark`                                      | boolean         | No       | Whether to add an AI-generated watermark.                                                           |
| `optimize_prompt_options.mode`                   | string          | No       | Prompt optimization mode. Seedream 5.0 Lite supports `standard`.                                    |

## Input limits

* Formats: `jpeg`, `png`, `webp`, `bmp`, `tiff`, `gif`, `heic`, and `heif`.
* Maximum file size: `30 MB` per image.
* Aspect ratio: `[1/16, 16]`.
* Width and height: each must be greater than `14px`.
* Total pixels: no more than `36,000,000` per image.
* Reference images: up to 14; the number of input references plus generated images must not exceed 15.

<Card title="API Reference" icon="code" href="/api-reference/model-api/bytedance/seedream-5-0-lite-260128-i2i">
  Open the interactive Seedream 5.0 Lite i2i API reference.
</Card>

<script src="/feedback.js" />