leeguooooo/image-use

What it solves

image-use is a zero-dependency Python CLI tool that allows users to generate images using their existing ChatGPT or Gemini subscriptions (including free tiers) without needing a separate OPENAI_API_KEY or paid API billing. It effectively turns a personal subscription into a programmatic interface for image generation.

How it works

The tool provides multiple backends to interact with AI services:

• Web Backend: Drives a logged-in Chrome browser to interact with the ChatGPT web chat interface.
• Codex Backend: A headless fallback that uses internal endpoints.
• Gemini Backends: Uses either a logged-in Chrome browser (--backend gemini) or the Antigravity CLI (--backend agy) to generate images via Google Gemini.

It supports advanced features like image-to-image reference, composition references, and the creation of simple animations by generating sprite sheets and processing them into WebP or GIF files using ImageMagick and libwebp.

Who it’s for

Developers and AI agent users who want to integrate image generation into their workflows or CLI environments (such as Claude Code, Cursor, or Codex) without paying for API credits, leveraging their existing web-based subscriptions instead.

Highlights

• No API Key Required: Uses existing browser sessions or internal endpoints to bypass the need for a funded API account.
• AI Agent Integration: Can be added as a "skill" to agents like Claude Code and Cursor via npx skills add.
• Character Consistency: Supports pinning characters via reference images to maintain consistency across different scenes.
• Community Styles: Integrates with a public gallery (drawstyle.leeguoo.com) to search, download, and publish custom art styles.
• Animation Support: Can generate ping-pong loop animations from sprite sheets.
• Multi-Backend Support: Flexible switching between ChatGPT and Gemini backends.