Image generation (Images)

Compatible with OpenAI’s Create image.

POST https://noviahub.com/v1/images/generations

For models whose endpoint labels include Image, such as gpt-image-2 (check on Models & Pricing).

Gemini image models don’t use this endpoint

Gemini image models such as gemini-3.1-flash-image cannot be called here. In our tests this returns 500: not supported model for image generation, only imagen models are supported. Use one of the methods under Gemini image models instead.

Headers

Header	Required	Description
Authorization	Yes	Bearer <your API key>
Content-Type	Yes	application/json

Request parameters

Parameter	Type	Required	Description
model	string	Yes	Model ID, such as gpt-image-2. Without it the gateway assumes dall-e, which NoviaHub doesn’t have, so the call returns 503.
prompt	string	Yes	Description of the image. The gateway doesn’t check it, but the model needs it.
n	integer	No	Number of images, 1 to 128. Missing or 0 means 1; above 128 returns an error.
size	string	No	Size, such as 1024x1024. Use the letter x, not the multiplication sign ×, or you get an error.
quality	string	No	Quality; values depend on the model.
response_format	string	No	How the image is returned, such as b64_json or url, if the model supports it.
background, output_format, output_compression, moderation, style, partial_images, input_fidelity	—	No	Forwarded as defined by OpenAI; support depends on the model.
stream	boolean	No	Stream the result (if the model supports it).
user	—	No	User identifier, forwarded as is.

Fields not in the table are not forwarded upstream.

Sample code

cURL

curl https://noviahub.com/v1/images/generations \
  -H "Authorization: Bearer $NOVIAHUB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "An orange cat sunbathing on a windowsill, watercolour style",
    "size": "1024x1024",
    "n": 1
  }'

Python

import base64
import os

from openai import OpenAI

client = OpenAI(
    base_url="https://noviahub.com/v1",
    api_key=os.environ["NOVIAHUB_API_KEY"],
)

result = client.images.generate(
    model="gpt-image-2",
    prompt="An orange cat sunbathing on a windowsill, watercolour style",
    size="1024x1024",
    n=1,
)

image = result.data[0]
if image.b64_json:
    with open("cat.png", "wb") as f:
        f.write(base64.b64decode(image.b64_json))
else:
    print(image.url)

TypeScript

import { writeFile } from 'node:fs/promises'

import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://noviahub.com/v1',
  apiKey: process.env.NOVIAHUB_API_KEY,
})

const result = await client.images.generate({
  model: 'gpt-image-2',
  prompt: 'An orange cat sunbathing on a windowsill, watercolour style',
  size: '1024x1024',
  n: 1,
})

const image = result.data?.[0]
if (image?.b64_json) {
  await writeFile('cat.png', Buffer.from(image.b64_json, 'base64'))
} else {
  console.log(image?.url)
}

Response

Field	Description
created	Creation time (Unix seconds).
data[]	One entry per image. The image is in b64_json (Base64) or url, depending on the model.
usage	Usage, when the model reports it.

Example response (test instance; image data shortened)

{
  "created": 1790609867,
  "data": [{ "b64_json": "iVBORw0KGgo=" }],
  "usage": { "input_tokens": 10, "output_tokens": 1056, "total_tokens": 1066 }
}

b64_json is the image file encoded in Base64; decode it and save it, as the samples do.

Whether a model is billed by tokens or per image, and at what price, is shown on its Models & Pricing page.

Gemini image models

Gemini image models such as gemini-3.1-flash-image can be called in two ways:

1. Native Gemini endpoint (recommended): POST /v1beta/models/gemini-3.1-flash-image:generateContent; see generateContent. The generated image is in candidates[].content.parts[].inlineData (mimeType plus Base64 data).
2. Chat Completions: POST /v1/chat/completions with this image model as model; the gateway asks the model for both text and image output. Images come back inside choices[0].message.content as Markdown images: ![image](data:<image type>;base64,<data>).
   • Set aspect ratio and size with extra_body.google.image_config, using the keys aspect_ratio and image_size (snake_case only; camelCase keys are rejected).

Common errors

HTTP status	Cause
400	n above 128 (n must be an integer between 1 and 128), the sign × in size, or the body isn’t valid JSON.
401	The key is invalid, disabled, expired or out of quota.
403	The key isn’t allowed to use this model, the IP isn’t on the allow list, or the account balance is too low.
500	A Gemini image model was called on this endpoint (convert_request_failed).
503	Wrong or missing model ID, or no channel is currently available for the model.

See Errors and troubleshooting for details.