Explore AI-generated images and prompts.
A working archive of prompt-driven photography, illustration, and product imagery — every piece traceable to the model that produced it and the exact text behind it. Browse the gallery, lift any prompt into the studio, remix it into your own work.
// Topics — what people are making
view all →// Engines — generated with
become pictures
Describe
Write what you want to see in plain language — subject, lighting, lens, mood. The closer to a photographer's brief, the better.
> studio portrait of an elderly fisherman, weathered hands, soft window light, 50mm
Engine
Pick a model. Each has a different fingerprint — Nano Banana renders cinematic, Pro is photographic, 2 leans illustrative.
engine = nano-banana-pro
ratio = 3:2
seed = 8847
Iterate
Remix from any image in this archive. The studio loads its prompt and model — change one variable at a time.
fork ./i/fisherman-port
- "weathered hands"
+ "calloused hands, wedding ring"
// Pick your engine
gpt-image-2
OpenAI's flagship image generation & editing model. Not just pretty pictures — structured commercial images with accurate text, editable workflows, and API control. If you're coming from DALL·E 3, this is the upgrade path.
nano-banana
The complete Nano Banana family in one workspace — text-to-image, image-to-image editing, Pro 4K renders, and the newest Nano Banana 2 reasoning engine, all on one page.
nano-banana-pro
Nano Banana Pro for production-grade 4K renders — sharp legible text, brand-quality detail, and optional reference images for character or style consistency.
Open editor 004 / Banananano-banana-2
Newest Nano Banana 2 with reasoning, real-world grounding, and 4K output — accepts up to 14 reference images and 20K-character prompts for the most ambitious generations.