Zhen AI / Image playground

Try GPT-Image text-to-image

Describe a picture and watch GPT-Image-2, 2.5 Flare or 2.5 Sunburst make it — right here, free, no signup. Add one reference image to restyle or edit it. Every finished image lands in the public gallery below, and the exact same request works against the Zhen AI API.

Text to imageNo signup required
Model
0/800
Reference image (optional)0/1

Add one image (JPEG/PNG/WebP, ≤10 MB) to restyle or edit — describe the change in your prompt. It is deleted from our storage after 24 hours.

Reference images of real people (including photorealistic AI portraits) aren't accepted on the free demo.

Size

The demo renders at a single default quality. Quality tiers are selectable through the API.

Free demo · 2 different original prompts per day, up to 800-character prompts

…

Your image appears here

Write a prompt (or pick an example) and hit Generate — no account needed.

1:1 · 1024x1024

A friendly note before you generate

This is a public demo. Anyone who tries to generate sexual content involving minors, or adult pornographic content, will permanently lose access to this site. Political and military subjects are not recommended here.

What we'd love to see: product shots, illustration, concept art and poster design.

Community gallery

Every image made in this playground, by everyone — yours will appear here too, even if you leave while it renders.

Nothing here yet — be the first to make an image.

No signup

Type a prompt and generate — no account, no card.

2 original prompts / 24h

Fair-use demo quota per visitor over a rolling 24-hour window — different prompts, not retakes of one. Near-repeats of a prompt you just ran aren't accepted on the free demo.

Same API under the hood

The playground calls the identical /v1/images endpoint your code would.

API: every size and quality

Sign up and add credits — choose quality tiers, any size and as many images as you need, billed per image.

Need more than 2 images a day, or want to pick the quality?

Sign up and add credits — the same three models through the API, with quality tiers, billed per image.