A Prompt-as-Code asset library that structures image-generation prompts into reusable, composable templates. It organizes 520 documented cases and supports GPT Image 2 workflows through a live browsing site and agent skill.
Project overview
A Prompt-as-Code asset library that structures image-generation prompts into reusable, composable templates. It organizes 520 documented cases and supports GPT Image 2 workflows through a live browsing site and agent skill.
Project type
AI Agent · Workflow · Image & Vision
Use cases
Image Processing
Deployment
Refer to project documentation
License
MIT
Best for
Creators who need to find and adapt documented image-generation prompt references from an organized gallery.
Developers building agent workflows that require selecting styles and templates from a shared library.
Teams managing batch or production image generation who need reusable, composable prompt templates.
Key capabilities
Provides reusable prompt templates that decompose visual elements into composable parts for controllable image-generation workflows.
Provides a gallery of 520 documented cases organized by visual category for finding prompt references.
Includes an agent skill that selects styles, templates, categories, and scene tags from the project's shared style-library data.
Offers a live website for browsing previews, copying complete prompts, filtering cases, and returning to source cases.
Limitations and risks
The homepage provides an English navigation layer, while the prompt body remains in the original template document.
The project references public prompt-library content owned by original authors or platforms, which may include infringing or inappropriate material subject to correction or removal.
GPU requirements, data boundaries, cost dependencies, coding requirements, and telemetry are not documented.
Getting started
Start from featured cases on the site and select your desired visual output type.
Browse the full gallery to find nearby cases relevant to your goal.
Input your business variables into the general or JSON templates to generate the final prompt.
README: This project turns scattered community examples into Prompt-as-Code assets that are easier for agents and automation workflows to reuse.
README: Browse all 520 cases by gallery part and category.
README: This repository includes an agent skill for choosing GPT-Image2 styles, templates, categories, and scene tags from the same data used by the website.
README: Use the live site at [gpt-image2.canghe.ai](https://gpt-image2.canghe.ai/) to browse the gallery as a product experience: open large previews, copy full prompts, filter by style o…