Start with the shown result
Open the material page and check the image, the exact prompt and the recorded model. Our current examples were generated in Google Flow. The model name shown on a page records that environment; it does not prove identical behaviour in Gemini Apps or another service.
Copy and set up
Use Copy prompt, then paste the text into your chosen image-generation interface. In Flow, choose image generation and the appropriate available model. Set portrait 3:4 for these examples using the interface controls where offered. This library does not start generation or buy credits for you. Check the provider’s current quota and any displayed charge before submitting.
Compare the first result
Compare subject, composition, material and light with the example. A new generation can differ even with the same text. Inspect edges, reflections, small objects and any accidental lettering at full size. A copied prompt is a starting point, not a guarantee of an identical output.
If copying fails
Some browsers restrict clipboard access. The full prompt remains selectable; use your device’s normal text-selection and copy controls. On a phone, you can also open the page in your usual browser.
Keep useful versions
Save the text that produced your own result along with the date, model and settings. If you change it, keep that version separately. This is more informative than remembering only the title of the example.
Further reading: Google DeepMind's image prompting guide.