Image2Studio
Turning visual prompting into reusable production scenes for real creative work.
- Expanding the GPT Image 2 prompt library
- Making outputs easier to discover and reuse
- Writing down what works
Hello, I’m Ryan—a developer in Shanghai exploring AI products, open-source tools, image-making, and the craft of shipping.
A small snapshot of what has my attention this month.
Turning visual prompting into reusable production scenes for real creative work.
AI makes creating fast. The interesting work now lives in publishing, checking, and turning repeated actions into durable tools.
Products and tools that each mark a part of my path.
Production-ready image scenes and a growing GPT Image 2 prompt library.
Private LLM inference entirely in the browser, powered by WebGPU.
Publish a static site from an agent session with one obvious command.
Browser checks and focused feedback for AI-built websites.
A live window into the code I have been touching recently.
A few commits that changed the direction.
Graphics, C++, OpenGL, and university experiments.
ChatLLM-Web became the first project to reach hundreds of developers.
Repeated workflows began turning into tools and reusable skills.
Creating products, publishing them, and sharing the systems behind the work.
The person continues after the code editor closes.
I collect interaction details and enjoy understanding why movement feels right.
I prefer a working artifact, a public URL, and lessons earned through use.
Working in UTC+8, following ideas wherever the internet carries them.