OpenAI Image Generator API

A FastAPI wrapper for OpenAI's image generation API that supports the new gpt-image-1 model as well as dall-e-2.

Features

• Generate images using OpenAI's gpt-image-1 model
• Edit images with multiple source images (supported by gpt-image-1 and dall-e-2)
• Create image variations (only supported by dall-e-2)
• Compatible with OpenAI's API parameter constraints
• Simple REST API interface with form data support for file uploads
• User-friendly web interface for all operations

Requirements

• Python 3.8+
• OpenAI API key

Installation

1. Clone this repository:

   git clone <repository-url>
   cd openai-image-generator
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Create a .env file from the example:

   cp .env.example .env
   

4. Add your OpenAI API key to the .env file:

   OPENAI_API_KEY=your_openai_api_key_here
   

Usage

1. Start the server:

   python main.py
   

   Alternatively, you can use uvicorn directly:

   uvicorn main:app --reload
   

2. Access the API documentation at http://localhost:8000/docs

API Endpoints

Generate Image

POST /images/generations

Parameters:

• prompt (string, required): A text description of the desired image
• model (string, default: "gpt-image-1"): The model to use for image generation
• n (integer, default: 1): The number of images to generate
• size (string, default: "1024x1024"): The size of the generated images ("1024x1024", "1792x1024", or "1024x1792")

Edit Image

POST /images/edits

Parameters:

• image (file(s), required): The image(s) to edit (multiple files allowed)
• prompt (string, required): A text description of the desired edited image
• model (string, default: "gpt-image-1"): The model to use for image editing
• n (integer, default: 1): The number of images to generate
• size (string, default: "1024x1024"): The size of the generated images

Create Image Variation

POST /images/variations

Parameters:

• image (file, required): The image to create variations of
• model (string, default: "dall-e-2"): The model to use (only "dall-e-2" is supported for variations)
• n (integer, default: 1): The number of variations to generate
• size (string, default: "1024x1024"): The size of the generated images

Examples

Using cURL

Generate an image with gpt-image-1:

curl -X 'POST' \
  'http://localhost:8000/images/generations' \
  -H 'accept: application/json' \
  -F 'prompt=A cute baby sea otter' \
  -F 'model=gpt-image-1' \
  -F 'n=1' \
  -F 'size=1024x1024'

Edit multiple images:

curl -X 'POST' \
  'http://localhost:8000/images/edits' \
  -H 'accept: application/json' \
  -F 'image=@body-lotion.png' \
  -F 'image=@bath-bomb.png' \
  -F 'image=@incense-kit.png' \
  -F 'image=@soap.png' \
  -F 'prompt=Create a lovely gift basket with these four items in it' \
  -F 'model=gpt-image-1'

Using Python with requests

import requests
import json
import base64

# Generate image
url = "http://localhost:8000/images/generations"
payload = {
    "prompt": "A cute baby sea otter",
    "model": "gpt-image-1",
    "n": 1,
    "size": "1024x1024"
}
response = requests.post(url, data=payload)
result = response.json()

# Save the image
with open("otter.png", "wb") as f:
    image_data = base64.b64decode(result["data"][0]["b64_json"])
    f.write(image_data)

Model Capabilities

gpt-image-1

• Image generation
• Image editing with multiple source images
• Supports sizes: 1024x1024, 1792x1024, 1024x1792

dall-e-2

• Image generation
• Image editing (single source image)
• Image variations
• Supports sizes: 256x256, 512x512, 1024x1024

Additional Notes

This application is designed to be compatible with the current version of the OpenAI API. OpenAI's API documentation mentions parameters like quality, style, and response_format, but these may not be supported in all versions of the API. If you encounter errors about unsupported parameters, you may need to update the OpenAI Python package.

License

MIT