Hi HN! I built a browser-based playground for Alibaba’s new Z-Image Turbo model and wanted to share it with the community.

Demo: https://z-img.net/

No GPU or setup required — everything runs through a hosted API. Key features:

Photorealistic image generation in a few seconds

Extremely accurate Chinese text rendering (rare among most models)

Simple, lightweight interface

Free to try; no login needed

I created this because many makers and developers want to experiment with modern image models without downloading huge weights or renting expensive GPUs. This tool aims to be a fast, frictionless sandbox.

I’d love feedback from the HN community on:

Latency / image quality

API design ideas

What features you think should exist (batch mode? upscaling? prompt templates?)

Any edge cases where the Chinese text handling breaks

Happy to answer questions and share technical details. Thanks for taking a look!