Skip to content
Projects

SnowSky1/universal-imagegen-agent

View on GitHub

SnowSky1/universal-imagegen-agent has 233 GitHub stars and 12 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-15 23:45 UTC. Last push 2026-10-11.

How these counts are measured

Measured star growth

in 24h
0
in 7 days
Collecting history
in 15 days
Collecting history
in 1 month
Collecting history

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

About

把 Codex imagegen 的工作流与提示词库带到任意智能体;支持 npm/pnpm 一键安装、配套 skill、独立 API 与 Qwen-Image 兼容网关。

AI

README.en.md

Original

Universal ImageGen Agent

Generate → edit walkthrough · Changelog · Author's personal showcase

Version 0.1.1 rejects n=0 consistently in the Node.js and Python CLIs. The default remains one image. The walkthrough includes a runnable local API fixture and separate instructions for a real provider.

Install the fixed version: v0.1.2 is published on npm through GitHub Actions trusted publishing. The GitHub Release also includes an npm archive, Python wheel, and source distribution.

npm install --global [email protected]

GitHub stars npm version npm downloads CI License

A runtime-neutral image generation and editing skill that brings the workflow, prompt engineering, use-case taxonomy, edit invariants, and prompt library of the OpenAI Codex imagegen skill to any general-purpose AI agent. Agents can reuse this mature image-generation methodology through a CLI or Python API without depending on a Codex-only built-in image tool.

This is especially useful as open image ecosystems such as Qwen-Image and Qwen-Image-2.1 grow. The project separates the agent-side workflow from the model backend: it can use OpenAI Images APIs, third-party compatible services, or Qwen-Image deployments exposed through an OpenAI-compatible gateway. A different serving protocol can be supported by adding a provider adapter without rewriting the prompt library or agent workflow.

The currently bundled provider is OpenAI-compatible. Native local Diffusers, ComfyUI, and SGLang backends for Qwen-Image are not built in yet.

It offers a CLI, a Python API, machine-readable JSON output, configurable credentials, custom base URLs, custom models, custom headers, and compatible Images API support.

The original Codex imagegen skill is preserved unchanged under upstream/imagegen/. The adapted root SKILL.md is suitable for agents with shell access and does not depend on a Codex-only built-in image tool.

Quick start

npm install --global universal-imagegen-agent
imagegen-agent install-skill
imagegen-agent config --json
imagegen-agent generate "A ceramic mug on a stone table" --dry-run --json

With pnpm:

pnpm add --global universal-imagegen-agent
imagegen-agent install-skill

Without a global install:

npx universal-imagegen-agent install-skill
pnpm dlx universal-imagegen-agent generate "A ceramic mug" --dry-run --json

Install the latest GitHub version before an npm release:

npm install --global github:SnowSky1/universal-imagegen-agent

For full documentation, see README.md.