Image Generation and Editing

Orbit AI API Team2026/07/116 min read

Orbit AI API provides OpenAI-compatible image endpoints for generation and editing. Point the Base URL to Orbit AI API and use an API key created in the console.

Connection parameters

Parameter	Value
Base URL	https://orbitaiapi.site/v1
Authentication	Authorization: Bearer YOUR_API_KEY
Model	Exact name shown in the console, such as gpt-image-2, gpt-image-2-all, or a Gemini image model

Image generation

Endpoint: POST /v1/images/generations

cURL example

curl -X POST "https://orbitaiapi.site/v1/images/generations" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A premium product poster with a clean background and realistic photography style",
    "size": "1024x1024",
    "quality": "auto",
    "n": 1,
    "response_format": "url"
  }'

Request fields

Field	Required	Description
model	Required	Image model name, such as gpt-image-2, gpt-image-2-all, or a Gemini image model.
prompt	Required	Description of the image to generate.
size	Optional	Common GPT sizes: 1024x1024, 1536x1024, 1024x1536. Gemini images may accept aspect ratios or model-specific sizes.
quality	Optional	Common values: auto, low, high, 2K, 4K. Support varies by model.
n	Optional	Number of images to generate. Usually 1.
response_format	Optional	url or b64_json. Prefer url in most cases.

Gemini image example

For Gemini image models, you can pass aspect ratio and resolution via size / quality, or provide Google-native options through extra_fields:

curl -X POST "https://orbitaiapi.site/v1/images/generations" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.1-flash-image",
    "prompt": "Generate a 9:16 cyberpunk character poster",
    "size": "9:16",
    "quality": "2K",
    "extra_fields": {
      "google": {
        "image_config": {
          "aspect_ratio": "9:16",
          "image_size": "2K"
        }
      },
      "aspect_ratio": "9:16",
      "image_size": "2K"
    },
    "response_format": "url"
  }'

JavaScript example

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.ORBIT_API_KEY,
  baseURL: 'https://orbitaiapi.site/v1',
});

const result = await client.images.generate({
  model: 'gpt-image-2',
  prompt: 'A premium product poster with a clean background and realistic photography style',
  size: '1024x1024',
  quality: 'auto',
  n: 1,
  response_format: 'url',
});

console.log(result.data[0].url);

Image editing

Endpoint: POST /v1/images/edits

Use this endpoint to modify an existing image based on a prompt. Requests typically use multipart/form-data to upload the source image.

cURL example

curl -X POST "https://orbitaiapi.site/v1/images/edits" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "model=gpt-image-2" \
  -F "image=@./input.png" \
  -F "prompt=Change the background to light gray while keeping the main product, realistic photography style" \
  -F "size=1024x1024" \
  -F "n=1" \
  -F "response_format=url"

Request fields

Field	Required	Description
model	Required	Image model that supports editing. Use the exact name from the console.
image	Required	Source image file. Common formats: PNG, JPG, WEBP.
prompt	Required	Edit instruction describing how the image should change.
mask	Optional	Mask image. White areas are editable; black areas are preserved. Supported by some models.
size	Optional	Output size such as 1024x1024, depending on model support.
n	Optional	Number of images to generate. Usually 1.
response_format	Optional	url or b64_json. Prefer url in most cases.
quality	Optional	Output quality. Common values: auto, low, high. Support varies by model.

JavaScript example

import fs from 'fs';
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.ORBIT_API_KEY,
  baseURL: 'https://orbitaiapi.site/v1',
});

const result = await client.images.edit({
  model: 'gpt-image-2',
  image: fs.createReadStream('./input.png'),
  prompt: 'Change the background to light gray while keeping the main product, realistic photography style',
  size: '1024x1024',
  n: 1,
  response_format: 'url',
});

console.log(result.data[0].url);

Response example

A successful response usually looks like this:

{
  "created": 1710000000,
  "data": [
    {
      "url": "https://example.com/generated-image.png"
    }
  ]
}

If you use response_format: "b64_json", the field is b64_json instead of url.

Best practices

• Copy model names from the Model list or console. Avoid hardcoding names that may be retired.
• For generation, write clear and specific prompts. For editing, state what to keep and what to change.
• Prefer response_format: "url" for easier preview and download.
• Call the API from your backend. Do not expose API keys in frontend code.
• Supported sizes, quality options, and editing capabilities vary by model. Always verify against the console and actual responses.

Common issues

• 401: Check that Authorization: Bearer YOUR_API_KEY is correct and the key is enabled.
• 404: Confirm the Base URL includes /v1 and the path is /images/generations or /images/edits.
• Model unavailable: Confirm account credit and use the exact model name shown in the console.
• Edit failed: Check image format and size requirements, and make the edit prompt more specific.

For general integration details, see OpenAI-compatible API.