Images

AI image generation, editing, variations, upscaling, style transfer, and background removal. Use SeaLink /v1/images/* and choose model capabilities and parameters from the model detail page; for your first test, use n=1 and a base size.

`POST /v1/images/generations`

AI image generation. The example model comes from the current public callable catalog; use /v1/models and the model detail page for available models, sizes, and reference-image support.

For image editing and reference-image models, data:image/...;base64,... inputs are the most reliable. If you pass a plain URL, it must be publicly accessible without sign-in, and both image dimensions must be greater than 10 px.

Parameters

Parameter	Type	Required	Description
model	string	Yes	gpt-image-2
prompt	string	Yes*	Image description; *no if using template_id
template_id	string	No	Prompt template ID
variables	object	No	Prompt template variables; values must be strings
n	integer	No	1 - 10 (default 1)
size	string	No	256x256, 512x512, 768x768, 1024x1024, 1328x1328, 1472x1104, 1104x1472, 1664x928, 928x1664, 1792x1024, 1024x1792, 1920x1920, 4096x4096. Seedream 4.0 accepts 1K, 2K, 4K, and WIDTHxHEIGHT from 921,600 to 16,777,216 total pixels; 2048x2048 is the default. Seedream 4.5 accepts 2K, 4K, and WIDTHxHEIGHT from 2560x1440 to 4096x4096 total pixels. Seedream 5.0 Lite accepts 2K, 3K, 4K, and WIDTHxHEIGHT from 2560x1440 to 4096x4096 total pixels. Dola Seedream 5.0 Pro accepts 1K, 2K, and WIDTHxHEIGHT from 921,600 to 4,624,220 total pixels; 2K is the default. For Gemini image models, use top-level aspect_ratio and image_size; delivered pixels follow the ratios and sizes supported by the selected model. See the model detail page for current pricing.
aspect_ratio	string	No	Model-specific. Gemini 3.1 Flash Image / Nano Banana 2 support 1:1, 1:4, 1:8, 2:3, 3:2, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 16:9, 21:9; 2.5 Flash Image, 3.1 Flash Lite Image, and 3 Pro Image support the common ratios and do not support 1:4, 1:8, 4:1, or 8:1. For Gemini image models, send this field at the top level, not inside response_format
image_size	string	No	Model-specific. Gemini 3.1 Flash Image / Nano Banana 2 support 512, 1K, 2K, 4K; 3 Pro Image / Nano Banana Pro support 1K, 2K, 4K; 3.1 Flash Lite Image / Nano Banana 2 Lite supports only 1K; 2.5 Flash Image uses a fixed 1K output tier and does not accept image_size
response_format	string	No	url, b64_json (default url). This only controls whether SeaLink returns a URL or base64; use top-level aspect_ratio and image_size for Gemini image models
quality	string	No	standard, hd
style	string	No	vivid, natural
quality	string	No	standard, hd
style	string	No	vivid, natural
image / image_url	string	No	Single reference image URL or base64 for models that support reference input; Seedream 4.0, 4.5, and 5.0 Lite accept up to 10 reference images, and reference images plus generated images must not exceed 15; one Seedream reference image can generate up to 14 images, and text-only image sets can generate up to 15 images; Dola Seedream 5.0 Pro accepts up to 10 reference images, the first reference image is free, and each additional reference image is billed; 2.5 Flash Image accepts up to 3 input/reference images total
images / image_urls / reference_images	string[]	No	Reference image array for models that support multiple references; use images: [url] when the selected model supports it; Seedream 4.0, 4.5, and 5.0 Lite accept up to 10 reference images, and reference images plus generated images must not exceed 15; one Seedream reference image can generate up to 14 images, and text-only image sets can generate up to 15 images; Dola Seedream 5.0 Pro accepts up to 10 reference images, the first reference image is free, and each additional reference image is billed; 2.5 Flash Image accepts up to 3 input/reference images total
reference_url / reference_urls	string / string[]	No	Reference image aliases; use only when supported by the selected model
source_lang / target_lang	string	No	Source and target language hints for image-translation models; supported values come from model details
domain_hint	string	No	Domain hint for image-translation OCR or layout work, such as general, doc, ecommerce, or comic
image_segment	boolean	No	Request segmentation assistance on supported image-translation models

cURL

curl https://api.sealink.io/v1/images/generations \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A serene Japanese garden with cherry blossoms, soft morning light",
    "n": 1,
    "size": "1024x1024",
    "response_format": "url"
  }'

Python

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.sealink.io/v1",
    api_key=os.environ["SEALINK_API_KEY"],
)

response = client.images.generate(
    model="gpt-image-2",
    prompt="A serene Japanese garden with cherry blossoms, soft morning light",
    n=1,
    size="1024x1024",
)
print(response.data[0].url)

Node.js

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.sealink.io/v1",
  apiKey: process.env.SEALINK_API_KEY || "",
});

const response = await client.images.generate({
  model: "gpt-image-2",
  prompt: "A serene Japanese garden with cherry blossoms, soft morning light",
  n: 1,
  size: "1024x1024",
});
console.log(response.data[0].url);

cURL - reference image

# Reference-image generation uses /v1/images/generations with images[]
IMG_B64="$(base64 -w0 photo.png)"
curl https://api.sealink.io/v1/images/generations \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<image-reference-model-from-/v1/models>",
    "prompt": "Turn this product photo into a clean catalog render",
    "images": ["data:image/png;base64,'"$IMG_B64"'"],
    "n": 1,
    "size": "1024x1024"
  }'

`POST /v1/images/edits`

Image editing. Choose a model whose detail page lists image editing support, then send it to /v1/images/edits. JSON and multipart/form-data are both supported; multipart files are limited to 10 MB.

Parameters

Parameter	Type	Required	Description
model	string	Yes	<image-edit-model-from-/v1/models>
prompt	string	Yes	Edit instruction
image / image_url	string	Yes	Source image base64 or URL; multipart mode uses image file field
mask / mask_url	string	No	Mask base64 or URL
n	integer	No	1 - 10
size	string	No	256x256, 512x512, 1024x1024
response_format	string	No	url, b64_json

cURL

# JSON mode — use a data URL for private or hard-to-fetch images
IMG_B64="$(base64 -w0 photo.png)"
curl https://api.sealink.io/v1/images/edits \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<image-edit-model-from-/v1/models>",
    "image_url": "data:image/png;base64,'"$IMG_B64"'",
    "prompt": "Add a rainbow in the sky",
    "n": 1,
    "size": "1024x1024"
  }'

Python

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.sealink.io/v1",
    api_key=os.environ["SEALINK_API_KEY"],
)

response = client.images.edit(
    model="<image-edit-model-from-/v1/models>",
    image=open("photo.png", "rb"),
    prompt="Add a rainbow in the sky",
    n=1,
    size="1024x1024",
)
print(response.data[0].url)

Node.js

import OpenAI from "openai";
import fs from "fs";

const client = new OpenAI({
  baseURL: "https://api.sealink.io/v1",
  apiKey: process.env.SEALINK_API_KEY || "",
});

const response = await client.images.edit({
  model: "<image-edit-model-from-/v1/models>",
  image: fs.createReadStream("photo.png"),
  prompt: "Add a rainbow in the sky",
  n: 1,
  size: "1024x1024",
});
console.log(response.data[0].url);

`POST /v1/images/variations`

Image variations. Generate variations of an input image with similar style but different content. Multipart/form-data only.

Parameters

Parameter	Type	Required	Description
image	file	Yes	PNG image file, max 10MB, square
model	string	No	<image-edit-model-from-/v1/models>
n	integer	No	1 - 10
size	string	No	256x256, 512x512, 1024x1024

cURL

# Multipart mode — upload local image
curl https://api.sealink.io/v1/images/variations \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -F "image=@photo.png" \
  -F "model=<image-edit-model-from-/v1/models>" \
  -F "n=1" \
  -F "size=1024x1024"

Python

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.sealink.io/v1",
    api_key=os.environ["SEALINK_API_KEY"],
)

response = client.images.create_variation(
    model="<image-edit-model-from-/v1/models>",
    image=open("photo.png", "rb"),
    n=1,
    size="1024x1024",
)
print(response.data[0].url)

Node.js

import OpenAI from "openai";
import fs from "fs";

const client = new OpenAI({
  baseURL: "https://api.sealink.io/v1",
  apiKey: process.env.SEALINK_API_KEY || "",
});

const response = await client.images.createVariation({
  model: "<image-edit-model-from-/v1/models>",
  image: fs.createReadStream("photo.png"),
  n: 1,
  size: "1024x1024",
});
console.log(response.data[0].url);

`POST /v1/images/upscale`

Image upscaling. Enlarge low-resolution images by 2x or 4x with enhanced clarity and detail.

Parameters

Parameter	Type	Required	Description
model	string	Yes	<image-upscale-model-from-/v1/models>
image	string	Yes	Source image base64 or URL
scale	integer	No	2 or 4
size	string	No	Optional output size or resolution label for metering on models that support it
quality	string	No	Optional quality tier for metering on models that support it
response_format	string	No	url, b64_json

cURL

curl https://api.sealink.io/v1/images/upscale \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<image-upscale-model-from-/v1/models>",
    "image": "https://example.com/low-res.jpg",
    "scale": 4,
    "response_format": "url"
  }'

Python

import os
import requests

res = requests.post(
    "https://api.sealink.io/v1/images/upscale",
    headers={"Authorization": f"Bearer {os.environ['SEALINK_API_KEY']}"},
    json={
        "model": "<image-upscale-model-from-/v1/models>",
        "image": "https://example.com/low-res.jpg",
        "scale": 4,
        "response_format": "url",
    },
)
print(res.json()["data"][0]["url"])

Node.js

const res = await fetch("https://api.sealink.io/v1/images/upscale", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.SEALINK_API_KEY || ""}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "<image-upscale-model-from-/v1/models>",
    image: "https://example.com/low-res.jpg",
    scale: 4,
    response_format: "url",
  }),
});
const data = await res.json();
console.log(data.data[0].url);

`POST /v1/images/style-transfer`

Style transfer. Apply the artistic style of one image to the content of another.

Parameters

Parameter	Type	Required	Description
model	string	Yes	<image-style-transfer-model-from-/v1/models>
image	string	Yes	Content image base64 or URL
style_reference	string	Yes	Style reference image base64 or URL
strength	number	No	0 - 1
size	string	No	Optional output size or resolution label for metering on models that support it
quality	string	No	Optional quality tier for metering on models that support it
response_format	string	No	url, b64_json

cURL

curl https://api.sealink.io/v1/images/style-transfer \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<image-style-transfer-model-from-/v1/models>",
    "image": "https://example.com/content.jpg",
    "style_reference": "https://example.com/style.jpg",
    "strength": 0.7,
    "response_format": "url"
  }'

Python

import os
import requests

res = requests.post(
    "https://api.sealink.io/v1/images/style-transfer",
    headers={"Authorization": f"Bearer {os.environ['SEALINK_API_KEY']}"},
    json={
        "model": "<image-style-transfer-model-from-/v1/models>",
        "image": "https://example.com/content.jpg",
        "style_reference": "https://example.com/style.jpg",
        "strength": 0.7,
        "response_format": "url",
    },
)
print(res.json()["data"][0]["url"])

Node.js

const res = await fetch("https://api.sealink.io/v1/images/style-transfer", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.SEALINK_API_KEY || ""}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "<image-style-transfer-model-from-/v1/models>",
    image: "https://example.com/content.jpg",
    style_reference: "https://example.com/style.jpg",
    strength: 0.7,
    response_format: "url",
  }),
});
const data = await res.json();
console.log(data.data[0].url);

`POST /v1/images/background-remove`

Background removal. Automatically detect and remove image backgrounds, returning a transparent PNG.

Parameters

Parameter	Type	Required	Description
model	string	Yes	<image-background-remove-model-from-/v1/models>
image	string	Yes	Source image base64 or URL
size	string	No	Optional output size or resolution label for metering on models that support it
quality	string	No	Optional quality tier for metering on models that support it
response_format	string	No	url, b64_json

cURL

curl https://api.sealink.io/v1/images/background-remove \
  -H "Authorization: Bearer $SEALINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<image-background-remove-model-from-/v1/models>",
    "image": "https://example.com/product.jpg",
    "response_format": "url"
  }'

Python

import os
import requests

res = requests.post(
    "https://api.sealink.io/v1/images/background-remove",
    headers={"Authorization": f"Bearer {os.environ['SEALINK_API_KEY']}"},
    json={
        "model": "<image-background-remove-model-from-/v1/models>",
        "image": "https://example.com/product.jpg",
        "response_format": "url",
    },
)
print(res.json()["data"][0]["url"])

Node.js

const res = await fetch("https://api.sealink.io/v1/images/background-remove", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.SEALINK_API_KEY || ""}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "<image-background-remove-model-from-/v1/models>",
    image: "https://example.com/product.jpg",
    response_format: "url",
  }),
});
const data = await res.json();
console.log(data.data[0].url);