Four-Part Story: Three Fluffy Creatures Build a Treehouse
Dettagli
| Categoria | Illustrazioni e anime |
|---|---|
| Formato | 16:9 |
| Input | Solo testo |
| Esempio generato con | GPT Image 2.5 |
| Modelli consigliati | Nano Banana 2, Seedream 5 Pro, GPT Image 2 |
| Autore originale | @GoogleAI · Post originale su X |
| Modello originale | nanobanana |
Perché questo prompt funziona
Instead of describing images, this prompt describes a story arc (a goal, a setback, teamwork, a happy ending) and leaves the visuals to the model. The one technical instruction that matters is 'maintain consistent identity across the 3 characters'. The author intended four separate 16:9 images generated one at a time, in a chat-style session. Run as a single generation, as in our example, the model drew all four beats in one frame: a bear, a bunny and a raccoon planning, failing ('Oops… not so easy'), working at sunset and celebrating in the finished treehouse. Either way, it is a fast route to a picture-book or social carousel draft. Name your creatures for tighter consistency.
Come adattarlo
| Variabile | Cosa cambia | Esempi |
|---|---|---|
| 3 fluffy creatures | the cast | a bear, a bunny and a raccoon, two penguins and a seal |
| building a treehouse | the shared goal | baking a giant cake, launching a paper boat |
| 4 images in 16:9 | panel count and format | 6 images in 1:1, 4 panels in one 16:9 image |
Consigli
- Name and describe each character once (species, colour, accessory) to keep them consistent across panels.
- For four separate images, generate in sequence and reuse the first image as a reference.
- Ask for short captions per panel if you want the story readable without words from you.
