GPT Image Studio

A small Python client for OpenAI GPT Image models with three entry points:

• CLI for terminal use and scripting
• PySide6 desktop GUI for local visual workflows
• NiceGUI web app for a polished browser interface

The shared generation logic lives in src/generator.py, so all three surfaces use the same OpenAI Image API wrapper. Selectable image models are configured in models.json.

Supported Models

Based on the current OpenAI image generation docs, the Image API supports these GPT Image models:

• gpt-image-2: latest state-of-the-art GPT Image model
• gpt-image-1.5: previous high-quality GPT Image model
• gpt-image-1: previous GPT Image model
• gpt-image-1-mini: cost-efficient GPT Image model

chatgpt-image-latest appears in OpenAI model listings, but it is not included here because this client calls client.images.generate() and focuses on Image API GPT Image models.

The Image API docs also mention legacy DALL·E models. dall-e-2 and dall-e-3 are deprecated, and their API support ended on May 12, 2026, so this client defaults to GPT Image models.

To add, remove, reorder, or change the default model, edit models.json. CLI validation and the desktop/web model dropdowns all read from that file.

Setup

uv sync
cp .env.example .env

Edit .env and set:

OPENAI_API_KEY=sk-your-api-key

Optional:

OPENAI_BASE_URL=https://your-proxy.example.com/v1
OPENAI_IMAGE_MODEL=gpt-image-2
IMAGE_OUTPUT_DIR=outputs

Leave OPENAI_BASE_URL unset when using the official OpenAI API. Set it only for OpenAI-compatible proxy providers or relay services, using the provider's API base URL.

OPENAI_IMAGE_MODEL overrides the default from models.json.

CLI

uv run python -m src.cli "A cinematic poster of an orange cat wearing an astronaut helmet"

Example with options:

uv run python -m src.cli \
  "A warm cyberpunk city at sunrise, cinematic composition" \
  --model gpt-image-2 \
  --size 1024x1024 \
  --quality high \
  --output-dir outputs \
  --count 1

Use --model to choose any model listed in models.json.

Desktop GUI

uv run python -m src.gui

Web App

uv run python -m src.web_app

Open:

http://127.0.0.1:5000

If port 5000 is already in use:

WEB_PORT=5001 uv run python -m src.web_app

Project Structure

src/
  config.py      # Environment and default configuration
  generator.py   # Shared OpenAI Image API wrapper
  models.py      # models.json loader
  cli.py         # CLI entry point
  gui.py         # PySide6 desktop GUI
  web_app.py     # NiceGUI web app
models.json      # Selectable image models and default model
outputs/         # Default image output directory, created at runtime

Notes

• GPT Image models may require OpenAI organization verification.
• Complex prompts may take longer to generate.
• Generated files are saved as PNG by default under outputs/.
• Dependencies are managed by uv through pyproject.toml and uv.lock.