Nano Banana Pro Prompts: 40 High-Fidelity Prompts for Portraits, Products and Print
Nano Banana Pro is the premium Gemini image model for complex instructions, high-fidelity product work, detailed layouts, real-world grounding and high-resolution visual production. These prompts are designed as briefs you can inspect, revise and hand to a human reviewer.
Updated September 28, 2026 · 42 practical prompts
The useful way to prompt an image model is to define what must remain stable, what should change, what the viewer needs to notice and how you will check the result. A long adjective pile is not a production brief. The templates below separate subject, action, setting, composition, light, constraints and review so you can adapt them to your own work.
First choose the right model and surface
Google’s current naming matters. Nano Banana Pro (Gemini 3 Pro Image) is an API model name, while “Nano Banana” is a product-family nickname. The Gemini app, Google AI Studio, Gemini API and Vertex AI may expose different controls, quotas and billing. Select the model shown by the product you are actually using, then confirm its current documentation before promising a resolution, reference count, grounding feature or price.
Use Pro when the brief is unusually demanding: a complex layout, exact brand treatment, detailed product surface, grounded visual explanation or high-resolution asset. That does not mean every job needs Pro. A simple crop, background swap or quick ideation pass may be cheaper and faster elsewhere.
A repeatable prompt structure
- Subject and reference: say what the attached image is and what must stay unchanged.
- Action: name one edit or generation goal instead of several competing changes.
- Composition: provide aspect ratio, camera position, crop, focal point and safe areas.
- Materials and light: describe surfaces, shadows, direction and colour temperature.
- Text: provide exact approved copy, or request a blank panel for later layout.
- Exclusions: list objects, claims, logos, identifiers or style constraints to avoid.
- Review: ask for a transcription, checklist or uncertainty report when appropriate.
For an edit, start with the invariant list: “keep the face, pose, clothing and crop unchanged.” For a blend, label the inputs: “Image A is the subject; Image B is the background.” For a product, separate the object from the scene: “preserve the package exactly; change only the surface and light.” These small sentences reduce accidental creative decisions.
How to run a safe image workflow
1. Prepare references
Use images you own or have permission to use. Remove unnecessary personal information, school names, addresses, badges and metadata before uploading. If a person is identifiable, get consent for the intended use. Keep a copy of the original and note the model, date, prompt and output version.
2. Start with one variable
Change the background, then the light, then the crop. If you ask for a new outfit, a new pose, a new age, a new location and a new visual style in one request, you will not know which instruction caused drift. One-variable edits also make revisions cheaper and easier to explain.
3. Inspect the result
Check identity, object count, hands, edges, perspective, labels, numbers, logos, reflections, shadows and unintended claims. Generated text can look plausible while being wrong. If a chart or package contains important copy, use the image as a layout concept and place final text with a controlled design workflow.
4. Ask for diagnosis, not applause
When an output misses the brief, ask the model to classify the problem as reference drift, text error, composition, lighting, geometry or missing constraint. Then make the smallest next edit. Keep the successful parts of the previous prompt; rewriting everything can discard useful constraints.
5. Keep provenance
Google documents SynthID on generated images. That is useful provenance, not a substitute for disclosure, consent, copyright review or fact checking. Store the source references, prompt, model ID, output size, editor decision and final approved asset together.
40 high-fidelity Nano Banana Pro prompts
1. Studio product hero
Prompt: Use the supplied product photography as the exact reference. Create a premium 4:5 hero image with the object centred on a warm stone plinth, controlled rim light, realistic materials and clean negative space. Preserve every product detail and add no claims.
Why it helps: Pro is useful when material and layout precision matter.
2. Complex packaging mockup
Prompt: Build a realistic shelf mockup from this approved packaging artwork. Keep every label, logo, colour, dieline and proportion exact. Use a quiet premium grocery shelf, believable reflections and no additional words or competing brands.
Why it helps: Use approved artwork as the source of truth.
3. Brand-consistent campaign
Prompt: Create three campaign scenes for the same supplied product: studio, home and outdoor. Preserve exact brand colours, package geometry and logo placement across all three. Return a consistency checklist with each image.
Why it helps: Complex campaigns need a review matrix.
4. Technical diagram
Prompt: Create a precise technical illustration from this verified component list and connection table. Use only supplied components, directional arrows and labels. Keep spacing consistent, add a legend, and leave unknown connections unresolved rather than guessing.
Why it helps: Do not turn a polished diagram into unverified documentation.
5. Data-heavy infographic
Prompt: Turn this approved table into a publication-ready infographic. Preserve every number, unit, date and category exactly. Use a clear hierarchy, accessible contrast and a separate source note area. If a value is ambiguous, flag it visibly.
Why it helps: Recalculate and proofread outside the image model.
6. Map visualisation
Prompt: Create a clean conceptual map from the supplied verified coordinates, labels and routes. Preserve relative relationships and mark the result as an illustration, not a navigation map. Do not invent roads, distances, borders or landmarks.
Why it helps: Grounding does not remove the need to check geography.
7. Editorial portrait
Prompt: Create a high-fidelity editorial portrait from this consenting adult reference. Preserve identity, age, skin texture and expression. Use a controlled three-quarter composition, soft key light, subtle background separation and no invented uniform, title or affiliation.
Why it helps: Higher fidelity does not make consent optional.
8. Multi-person group
Prompt: Use these five consenting adult references to create one natural team portrait. Keep each face recognisable, preserve relative heights and avoid merging features. Use a neutral studio, balanced spacing, consistent light and no invented company logo.
Why it helps: Check each person separately at full size.
9. Product cutaway
Prompt: Create a conceptual cutaway of this supplied device. Show only the internal parts listed below, label them with the exact supplied names and mark the image 'conceptual'. Do not imply the drawing is an engineering specification.
Why it helps: A cutaway can communicate structure without certifying it.
10. Print poster
Prompt: Create a print-oriented poster at a portrait ratio with the exact approved headline, subhead and date. Use a restrained grid, strong typographic hierarchy and generous margins. Render no extra words; return a transcription and uncertainty list.
Why it helps: Prepare final print text in a layout tool when possible.
11. Scientific explainer
Prompt: Illustrate this supplied scientific process for a general audience. Keep the sequence and labels exact, separate observation from interpretation, use a clean cross-section and include a blank citation footer for the editor.
Why it helps: Visual clarity cannot replace source review.
12. Real-world grounding brief
Prompt: Using the supplied verified facts and current reference links, create a factual visual summary. Keep uncertain facts out of the image, use a clearly marked 'as of' date and return a fact-to-element mapping for review.
Why it helps: Grounded generation still needs an editorial audit.
13. Premium interior
Prompt: Redesign this interior concept while preserving the room's architecture, dimensions, windows and camera position. Replace only the materials and furniture listed in the brief. Show believable scale, joinery and light; do not claim buildability.
Why it helps: Use Pro for detailed concept exploration, not approval.
14. Architecture presentation
Prompt: Create a cohesive three-view presentation board from this architectural brief: exterior, plan-like diagram and interior. Use the same material palette, orientation and project label; keep the board labelled 'concept study' and leave the title block editable.
Why it helps: Consistency across views is the real task.
15. Luxury still life
Prompt: Create a high-fidelity still life from these three supplied objects. Preserve their exact shapes and finishes, arrange them in a calm editorial composition, use directional light and realistic contact shadows, and add no brand claims or text.
Why it helps: Ask for the physical relationships you need.
16. Fashion detail
Prompt: Use this garment reference to create a close editorial detail showing weave, seam, button and drape. Preserve the actual construction and colour. Do not add luxury, sustainability, origin or performance claims absent from the brief.
Why it helps: Material detail is not evidence of manufacturing facts.
17. Food packaging
Prompt: Create a premium food-packaging concept around the supplied dieline and approved product name. Keep all legal copy blank except the supplied name. Use a realistic material finish and return a separate list of areas requiring regulatory review.
Why it helps: Keep legal copy out of generative uncertainty.
18. Consistent character sheet
Prompt: Create a character reference sheet with front, side and three-quarter views from this approved design. Preserve face, clothing, colours and accessories. Use neutral poses, equal scale and labels View A, View B and View C only.
Why it helps: Reference sheets reduce drift in later scenes.
19. Film keyframe board
Prompt: Create six storyboard keyframes from the supplied shot list. Preserve the protagonist, wardrobe, prop placement and screen direction across frames. Use short shot numbers only; do not invent dialogue or continuity.
Why it helps: The shot list remains the authority.
20. Thumbnail system
Prompt: Create a reusable thumbnail system with three layouts: subject left, subject right and centred. Use one approved colour palette, safe margins and blank headline areas. Return the design rules separately so editors can apply them consistently.
Why it helps: Templates are more durable than one-off images.
21. Brand colour audit
Prompt: Compare this generated concept with the supplied brand palette. Return the image plus a swatch table showing intended, visible and uncertain colours. Do not claim exact colour values unless they were supplied or measured.
Why it helps: Ask the model to surface uncertainty.
22. High-resolution label
Prompt: Create a 4K product scene with a blank label panel precisely sized for the approved artwork. Preserve perspective, reflections and material edges, but do not generate small legal text. Keep the panel clean for later compositing.
Why it helps: Blank space often beats unreliable fine print.
23. Poster localisation
Prompt: Create localised poster layouts for English, French and Portuguese using the supplied translations. Preserve names, dates and URLs, allow for text expansion and return each text block separately. Do not translate legal names without approval.
Why it helps: Design for language expansion, not literal fitting.
24. Comparative product board
Prompt: Create a premium comparison board for products A, B and C using only the supplied objects and blank feature rows. Keep equal scale, neutral lighting and a consistent angle. Do not infer which product is better.
Why it helps: Comparison claims belong in verified copy.
25. Cinematic environment
Prompt: Create a cinematic environment concept from this location brief: preserve the stated architecture, time period and practical light sources, but avoid recognisable real brands and unverified historical detail. Return a continuity checklist.
Why it helps: Specific atmosphere can coexist with factual restraint.
26. Medical illustration concept
Prompt: Create an educational conceptual diagram from the supplied anatomy brief. Use neutral colours, clear labels and a prominent 'educational illustration' note. Do not diagnose, personalise, or imply that the visual replaces clinical advice.
Why it helps: High-fidelity visuals need stronger safety language.
27. Financial explainer
Prompt: Create a neutral educational visual from these verified figures. Preserve currency, dates and units exactly, include a blank source footer and avoid predictions, promises or personalised advice. Return a number-by-number verification checklist.
Why it helps: Never let visual polish imply financial certainty.
28. World knowledge check
Prompt: Before creating this image, list which factual elements require current verification and which can be treated as visual invention. Use only the verified elements in the final image and mark the rest as fictional or blank.
Why it helps: Separate research from composition.
29. Object fidelity
Prompt: Use these six reference images of one object to create a single product hero. Reconcile only details visible consistently across the references, list conflicts, preserve the most conservative interpretation and do not invent hidden surfaces.
Why it helps: Multiple references should expose conflicts, not hide them.
30. Editorial collage
Prompt: Create a sophisticated editorial collage from the supplied photos and shapes. Keep each supplied subject recognisable, preserve crop intent, avoid altering faces and leave captions blank. Use a single visual rhythm and generous margins.
Why it helps: Collage direction should specify what must not change.
31. Museum-style object plate
Prompt: Create a museum-style object plate from this artefact reference: neutral background, even light, front and detail views, no invented provenance or date. Leave the catalogue fields blank for a curator to complete.
Why it helps: Do not manufacture historical context.
32. Retail window
Prompt: Create a retail-window concept using the supplied products, approved seasonal palette and exact logo artwork. Keep all products at believable scale and leave promotional copy blank. Return a placement plan for the merchandiser.
Why it helps: The model can propose composition, not commercial claims.
33. Fine-art composition
Prompt: Create a gallery-style composition from these owned still-life references using a muted palette, controlled light and deliberate negative space. Preserve the objects' identity, avoid copying a named living artist and return the visual decisions separately.
Why it helps: Describe craft choices instead of requesting imitation.
34. Accessible chart poster
Prompt: Create a high-contrast chart poster from the supplied categories and values. Use direct labels, a texture or shape distinction in addition to colour, large type and a blank source footer. Preserve every value exactly.
Why it helps: Accessibility should be specified, then checked.
35. Instruction card
Prompt: Create a six-step instruction card from these verified steps. Use one action per panel, consistent icons, numbered sequence and no extra safety advice. Leave a large blank area for a human editor to add checked warnings.
Why it helps: Do not let the model invent safety instructions.
36. Before-and-after layout
Prompt: Create a before-and-after layout using the supplied original and approved edited image. Keep both images unchanged, label them Before and After only, use equal crops and do not imply the edit proves a performance or health outcome.
Why it helps: Show the source alongside the edit.
37. Scene continuity
Prompt: Create three sequential frames from this shot brief. Preserve camera direction, subject identity, prop position and time of day. Change only the stated action, and return a continuity table listing each invariant.
Why it helps: Continuity tables make drift visible.
38. Production handoff
Prompt: Turn this visual brief into a production handoff image and structured checklist: subject, dimensions, assets, copy, exclusions, open questions and approval owner. Keep open questions visible and do not fill them with guesses.
Why it helps: A good handoff preserves uncertainty.
39. Image-search grounded concept
Prompt: Use image search grounding only for the supplied subject category, not identifiable private people. Show the factual references used, avoid copying a source image, and create an original concept with a separate attribution and review checklist.
Why it helps: Follow Google’s display and attribution requirements when grounding.
40. Four-resolution export plan
Prompt: Create one composition designed for 0.5K, 1K, 2K and 4K exports. Keep the focal subject inside all safe areas, avoid tiny essential text and return an export checklist with intended use for each size.
Why it helps: Resolution choice should follow the delivery need.
41. Prompt-to-brief converter
Prompt: Convert this vague request into a professional visual brief with explicit subject, references, composition, materials, lighting, text, exclusions, output size and review gates. Do not add creative details unless you label them as optional.
Why it helps: The first job is to make the request testable.
42. Failure diagnosis
Prompt: Compare this output with the brief and classify each problem as reference drift, text error, composition, lighting, geometry, missing constraint or unsupported claim. Suggest the smallest next prompt change for each problem.
Why it helps: Small corrective prompts preserve useful work.
Prompt adaptation examples
For a thumbnail, keep the subject simple and reserve text space: “show the clearest moment, one focal point, high contrast, 16:9, empty right third for approved headline.” For a catalogue image, lock the object: “same product, exact shape and label placement, neutral surface, soft shadow, no claims.” For an infographic, do the opposite of improvisation: supply the verified labels, numbers and units, then ask the model to leave unclear items blank.
For a portrait, define dignity and identity: “preserve the consenting adult’s actual face, age, skin texture and body shape; change only the background.” For a multi-image composition, state which reference controls which property. For a series, repeat the invariant list in every prompt and compare the outputs side by side. Consistency is a process, not a single magic phrase.
Limitations worth planning for
- Exact identity, geometry and small text can drift even when the prompt is specific.
- A generated image may add plausible objects, reflections or claims that were never supplied.
- App and API controls are not automatically identical; access, quotas and prices can change.
- Grounding can help with current context, but sourced facts still need editorial checking.
- Image generation is not a substitute for engineering, medical, legal, financial or accessibility review.
- Do not upload private or sensitive images unless the people involved understand and consent to the workflow.
FAQ
Is Nano Banana 2 the same as Nano Banana Pro?
No. Google documents them as different models: Nano Banana 2 is Gemini 3.1 Flash Image and Nano Banana Pro is Gemini 3 Pro Image. Their controls, positioning and pricing can differ.
Can these prompts guarantee readable text?
No. They can constrain the request and ask for a transcription, but important copy should be checked against an approved source and often composited outside the model.
Can I use these prompts in the Gemini app?
They are written as portable briefs. The app, AI Studio, API and Vertex AI may expose different upload, model, size, grounding and billing controls. Adapt the prompt to the surface you are using.
How should I compare outputs?
Define acceptance criteria before generation: identity preserved, object exact, text correct, composition usable, no unsupported claims and suitable output size. Compare against those criteria rather than choosing the most attractive image by instinct alone.
For a broader library, see Nano Banana prompts, Gemini image prompts and AI image prompts.