Images

Image generation

Create images from a text prompt with an OpenAI-compatible endpoint.

POST /v1/images/generations

Generates one or more images from a text prompt. The endpoint follows the OpenAI Images API, so the official OpenAI SDKs work unchanged — point them at https://rutoken.ai/v1. Models such as gpt-image-1, dall-e-3, dall-e-2 and Google Imagen (imagen-*) can be called through it; the exact list depends on what is enabled in the Model Square.

Authentication

Send your key in the Authorization: Bearer sk-xxxx header.

Request body

Parameter	Type	Required	Description
`model`	string	Yes	Model ID, e.g. `gpt-image-1`, `dall-e-3`.
`prompt`	string	Yes	Text description of the image. Max length: 32000 characters for `gpt-image-1`, 4000 for `dall-e-3`, 1000 for `dall-e-2`.
`n`	integer	No	Number of images, 1–10. `dall-e-3` supports only `1`.
`size`	string	No	`gpt-image-1`: `1024x1024`, `1536x1024`, `1024x1536` or `auto`. `dall-e-3`: `1024x1024`, `1792x1024`, `1024x1792`. `dall-e-2`: `256x256`, `512x512`, `1024x1024`. Use the letter `x`, not `×`.
`quality`	string	No	`dall-e-3`: `standard` (default) or `hd`. `gpt-image-1`: `low`, `medium`, `high` or `auto`.
`style`	string	No	`dall-e-3` only: `vivid` or `natural`.
`response_format`	string	No	`url` or `b64_json`. `gpt-image-1` always returns `b64_json`.
`background`	string	No	`gpt-image-1` only: `transparent`, `opaque` or `auto`. A transparent background needs `output_format` `png` or `webp`.
`output_format`	string	No	`gpt-image-1` only: `png`, `jpeg` or `webp`.
`output_compression`	integer	No	`gpt-image-1` only, 0–100, for `jpeg` / `webp`.
`moderation`	string	No	`gpt-image-1` only: `low` or `auto`.
`stream`	boolean	No	Stream partial images (`gpt-image-1`, when the upstream supports it).
`partial_images`	integer	No	Number of partial images to stream, 0–3.
`user`	string	No	End-user identifier for abuse monitoring.

When the request is routed to a Google Imagen model, size is converted to an aspect ratio (1792x1024 → 16:9, 1024x1536 → 2:3, …). You can also pass a ratio such as 16:9 directly.

Request example

cURL

curl https://rutoken.ai/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $API_KEY" \
  -d '{
    "model": "gpt-image-1",
    "prompt": "A watercolor fox reading a book under a tree",
    "size": "1024x1024",
    "quality": "medium"
  }'

Python

import base64
from openai import OpenAI

client = OpenAI(base_url="https://rutoken.ai/v1", api_key="sk-xxxx")
resp = client.images.generate(
    model="gpt-image-1",
    prompt="A watercolor fox reading a book under a tree",
    size="1024x1024",
    quality="medium",
)
# gpt-image-1 returns base64 data
with open("fox.png", "wb") as f:
    f.write(base64.b64decode(resp.data[0].b64_json))

Node.js

import fs from "node:fs";
import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://rutoken.ai/v1", apiKey: "sk-xxxx" });
const resp = await client.images.generate({
  model: "gpt-image-1",
  prompt: "A watercolor fox reading a book under a tree",
  size: "1024x1024",
  quality: "medium",
});
// gpt-image-1 returns base64 data
fs.writeFileSync("fox.png", Buffer.from(resp.data[0].b64_json, "base64"));

A dall-e-3 request that returns a URL:

cURL

curl https://rutoken.ai/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $API_KEY" \
  -d '{
    "model": "dall-e-3",
    "prompt": "A futuristic city skyline at sunset, cinematic lighting",
    "size": "1792x1024",
    "quality": "hd",
    "style": "vivid",
    "response_format": "url"
  }'

Response

json

{
  "created": 1760000000,
  "data": [
    {
      "url": "https://example.com/generated/image.png",
      "revised_prompt": "A futuristic city skyline at sunset with cinematic lighting..."
    }
  ],
  "usage": {
    "total_tokens": 4240,
    "input_tokens": 50,
    "output_tokens": 4190,
    "input_tokens_details": { "text_tokens": 50, "image_tokens": 0 }
  }
}

Field	Type	Description
`created`	integer	Unix timestamp.
`data[].url`	string	Image URL (when `response_format` is `url`). Upstream URLs usually expire after a short time — download them promptly.
`data[].b64_json`	string	Base64-encoded image (when `response_format` is `b64_json`, and always for `gpt-image-1`).
`data[].revised_prompt`	string	Prompt actually used by `dall-e-3`.
`usage`	object	Token usage, returned by token-billed models such as `gpt-image-1`.

Gemini image models

Gemini's native image models (for example gemini-2.5-flash-image) are not served by this endpoint — only Imagen models are. Call them through:

Chat Completions — POST /v1/chat/completions with the image model as model;
the Gemini native format — POST /v1beta/models/{model}:generateContent with x-goog-api-key: sk-xxxx.