Free Nano Banana Prompts to Copy and Paste
The viral figurine trend, one-prompt photo edits, product mockups — this gallery pairs public creator results labelled as Nano Banana with the prompts shared in their source posts. Copy a prompt and adapt it in Gemini or Google AI Studio.
Nano Banana (Google's Gemini image model) is exceptional at editing an existing photo from a single instruction, which is why most prompts here are written around a reference image: upload your photo, paste the prompt, and the style transfers while your subject stays intact.
The collection is sourced from public creator posts and screened for prompt completeness, source attribution and engagement signals. Those signals help curation but do not guarantee the same output in another model version or account.
A copy-and-paste Nano Banana prompt structure
The most reusable examples read like a short creative brief rather than a pile of adjectives. Start with the uploaded reference and the desired outcome, then separate what must stay from what should change. Finish with the visual treatment and a small set of constraints so the prompt remains useful when you swap in another subject, photo or product.
- Reference and subject: identify the uploaded photo, person, object or product, including the detail the model should use as its anchor.
- Keep: name identity, pose, facial features, product shape, label placement or other details that must remain recognizable.
- Change: state the edit directly — restyle the outfit, move the scene, turn the subject into a figurine, or redesign the background.
- Style, camera and light: add the medium, framing, lens feel, color palette and lighting direction that define the finished image.
- Constraints: specify a clean composition, readable short text, natural hands or another high-priority check without overloading the request.
Choose an example by the job
Use the prompt collection as a library of starting points. Find an example with the same kind of reference and desired change, then replace only the variable details first. This preserves the useful structure while making the result your own instead of copying an unrelated style recipe into every task.
- Photo restyle: keep the subject and pose, then replace the medium, setting, wardrobe or color grade for a controlled transformation.
- Portrait and identity-preserving edits: describe the new light or styling while explicitly preserving the person’s recognizable features. Edit a real person’s likeness only with that person’s permission.
- Figurine or character scene: retain the reference subject and add the material, scale, packaging, props and studio setup one block at a time.
- Product photography: keep the product geometry and label details, then vary the surface, camera angle, light and commercial setting.
Fix common Nano Banana editing failures
When an edit misses the brief, shorten the instruction and make the conflict visible. Give the model one priority at a time: first protect the reference subject, then introduce the transformation, and finally add styling. The public examples are useful because they reveal which details creators chose to state explicitly.
- Identity drift: move “preserve the person’s facial features and proportions” near the reference instruction and remove competing changes to the face.
- Background carryover: describe the new background as a complete scene and say which old background elements must be removed.
- Garbled text: quote only the short wording that must appear, specify its location, and plan to verify every character in the result.
- Overconstraint: remove low-priority adjectives and keep the subject, transformation, composition and one lighting direction as the core request.
How this prompt collection is curated
PicGens organizes examples that are publicly posted with a visible result, a shared prompt and source attribution. We look for enough prompt text to explain the image, keep the model label attached to the card, and prefer recipes that can be adapted to a different subject or reference. A public post is a source for study, not a guarantee of permission, licensing or identical reproduction; review the creator’s terms before using any prompt or image in a project.
- Read the source context before adapting an example, especially when it includes a person, a brand, a private reference photo or recognizable artwork.
- Keep the structure that explains the result, but replace names, subjects and visual details with material you have permission to use.
- Treat model version, account settings and reference quality as variables; the same prompt can produce a different image on another run.
Frequently asked questions
- Where do I run Nano Banana prompts?
- In the Gemini app or gemini.google.com (choose image generation), or Google AI Studio. Copy a prompt from this page, optionally attach your own photo, and send.
- What is the difference between Nano Banana, Nano Banana Pro and Nano Banana 2?
- They are successive versions of Google's Gemini image model — later versions follow instructions better and handle text and multi-image consistency more reliably. Each card retains the version label supplied with its source, and we keep a separate gallery for Nano Banana 2. Results vary by version and account.
- Why do so many Nano Banana prompts mention uploading a photo?
- Because editing is the model's superpower: a single instruction like "turn this photo into a collectible figurine in blister packaging" restyles your image. Prompts written for editing produce far more personal, shareable results than pure text-to-image.