What is this?
Yingzao is a Claude Code / Codex skill that turns photos of Chinese architecture, cultural sites, shops and travel shots into art-directed editorial posters generated by GPT Image, instead of applying a filter or pasting a title over blank space.
Yingzao is a Claude Code / Codex skill that turns photos of Chinese architecture, cultural sites, shops and travel shots into art-directed editorial posters generated by GPT Image, instead of applying a filter or pasting a title over blank space.Is this a standalone app or something that runs inside an agent?It is an Agent Skill, not a standalone app — it only works inside a Claude Code or Codex agent that can read local images and call an image generation/editing tool.
It is an Agent Skill, not a standalone app — it only works inside a Claude Code or Codex agent that can read local images and call an image generation/editing tool.How do I start using it?After installing the skill, you just tell your Agent in natural language, inside a Skills- and image-generation-capable environment, what poster you want from your photo.
After installing the skill, you just tell your Agent in natural language, inside a Skills- and image-generation-capable environment, what poster you want from your photo.What do I need installed for it to work?You need Python 3.10+ with Pillow, NumPy, OpenCV and fontTools installed, plus an agent like Claude Code or Codex that can call an image-generation tool such as GPT Image.
You need Python 3.10+ with Pillow, NumPy, OpenCV and fontTools installed, plus an agent like Claude Code or Codex that can call an image-generation tool such as GPT Image.What is it not meant for?It is not built for ordinary product/e-commerce shots, simple fixes like exposure or sky swaps, invented ancient buildings with no real location, or jobs that only need a deterministic collage with text stuck on top.
It is not built for ordinary product/e-commerce shots, simple fixes like exposure or sky swaps, invented ancient buildings with no real location, or jobs that only need a deterministic collage with text stuck on top.What happens right after the poster is generated?Yingzao does a single low-cost re-read of the finished poster to check whether the subject treatment, text legibility and layout actually came through, and flags the clearest problem — but it never auto-regenerates or spends another image credit unless you ask for a fix.
Yingzao does a single low-cost re-read of the finished poster to check whether the subject treatment, text legibility and layout actually came through, and flags the clearest problem — but it never auto-regenerates or spends another image credit unless you ask for a fix.Can it combine several photos into one poster?Yes — in its 'fusion' mode it extracts the subject from each photo and rebuilds one scene with shared perspective, light direction and shadows, only keeping separate photo rectangles if you explicitly ask for a grid layout.
Yes — in its 'fusion' mode it extracts the subject from each photo and rebuilds one scene with shared perspective, light direction and shadows, only keeping separate photo rectangles if you explicitly ask for a grid layout.Why does it design the Chinese title text instead of just dropping in a font?It engineers the display characters' width, stroke contrast, closing strokes and rhythm, pairing smaller text with a different regular typeface for hierarchy, so the title reads as part of the artwork rather than a label pasted on top.
It engineers the display characters' width, stroke contrast, closing strokes and rhythm, pairing smaller text with a different regular typeface for hierarchy, so the title reads as part of the artwork rather than a label pasted on top.What are similar projects?mono-color-skill, guizang-social-card-skill, travel-memory-sticker-card, photo-abstract-editorial, gpt-image-skill, shuohao-skills, sprite-gen, GPT-Image2-Skill
mono-color-skill, guizang-social-card-skill, travel-memory-sticker-card, photo-abstract-editorial, gpt-image-skill, shuohao-skills, sprite-gen, GPT-Image2-SkillWhich topics is it in?gpt-image, photography, xiaohongshu, claude-skill, image-editing, content-creation, agent-skill, image-generation, image-processing, skill, skills, generative-ai
gpt-image, photography, xiaohongshu, claude-skill, image-editing, content-creation, agent-skill, image-generation, image-processing, skill, skills, generative-ai