Nano Banana Pro
from 25 crStudio-grade images with sharp text and lighting, up to 4K.
- 1K · 2K · 4K
- Up to 8 reference images
Whisk AI remixes up to three pictures into one: drop a subject, a scene and a style, add a few words if you like, and Nano Banana or GPT Image blends them. Independent tool, not affiliated with Google.
Done remixing? Animate it in Whisk AI VideoFree-plan results appear in the public gallery · charged only when it succeeds

Your image appears here
Pick a model, write a prompt and press Generate. You only pay for successful renders.
Whisk-style image remixer
WhiskAI is an independent Whisk AI–style image remixer. Instead of describing everything in words, you show it: one picture for the subject (a pet, a product, a character), one for the scene (a beach, a kitchen, a neon street) and one for the style (a sticker sheet, a watercolor, a clay toy). Every slot is optional, and so is the prompt. You can add up to 16 reference images in total when a model supports it.
The remix runs on Nano Banana (multi-reference) or GPT Image through image-to-image, up to 4K. You see the credit cost before you generate, failed images are refunded automatically, and every result has an "Animate this" button that opens the video studio with your image as the first frame. WhiskAI is not affiliated with Google; Whisk and Google are trademarks of Google LLC.
Upload subject, scene and style
Fill the slots you need, skip the rest. Images start at 6 credits.
The thing that must stay recognisable: your dog, a mascot, a bottle, a sketch of a character. A clear photo on a plain background works best.
The scene sets the place and composition; the style sets the look — medium, palette, texture. Leave either empty and the model fills it in. You can add up to 16 references in total.
A few words steer the blend: "as an enamel pin", "make it winter", "add the title SUMMER CLUB". With no prompt at all, the model simply combines the roles.
Download the image, reuse the same slots with a new prompt, run it again for a variation, or press Animate this to turn it into a clip.
Choose a model
Nano Banana keeps faces and products faithful across several references; GPT Image leans stylised with crisp lettering. Resolutions, reference limits and prices come straight from the studio.
Studio-grade images with sharp text and lighting, up to 4K.
Fast, faithful edits that keep faces and products consistent.
Detailed renders with wide aspect ratios such as 21:9.
Quick drafts and variations at the lowest price.
The original Nano Banana, still great for simple photo edits.
Vivid, stylized images with crisp typography for posters and covers.
Warm, photographic looks tuned for realism.
Dependable layouts, product shots and infographics up to 4K.
Low-cost text-to-image for brainstorming ideas.
Strong prompt following with up to 16 reference images.
The classic GPT-4o image model for friendly, illustrative results.
Example remixes
Start from words, from your own photos, or from a product. Open any example to reuse its prompt.
No uploads at all: describe the subject, place and look, and the studio works as a regular text-to-image generator.
Put a portrait or pet in the Subject slot and a place in the Scene slot; the remix moves the subject into that setting and matches the light.
Use a product shot as the subject and a mood board as the style: the shape and label stay, the look changes. Up to 16 references.
Use cases
Drop the product in as the subject and try ten scenes before you book a shoot.
Event and game posters with a consistent character and readable titles.
Your photo as the subject, a painting as the style: profile pictures in any look.
Turn a room photo into a cozy scrapbook poster or a cutaway miniature by pairing it with a style reference.
Explore wordmarks and mascots by remixing a rough sketch with a style reference.
Put a design on stickers, pins, tees or plush toys to preview merch.
Community
Real prompts and results from the community. Open any piece to read its prompt, then recreate it in the Whisk AI studio.
Create a 3D isometric diorama illustration of an original character named [CHARACTER NAME], peacefully working or practicing a creative activity in a nature-inspired environment, sitting at a wooden desk near a large…
Surrounding the realistic main subject are multiple cute, 3D-style chibi miniatures of the same person, with identical facial features, hairstyle, body proportions, and outfit. The chibi figures are naturally…
An isometric 3D cube-shaped miniature room (shallow cutaway true cube; everything strictly contained within the cube). The room is [Room Description]. Character: a chibi/figurine-style — [Character Description]. The…
2x2 grid, 16:9, do this for 4 famous cities: M₁ PERSONIFIED ICON MASCOT (master) the place's most iconic figure/monument personified as a friendly mascot character (face, hair, pose), central and dominant. ¬M₁ = no…
A high quality mixed media portrait of a trendy young man (using my uploaded person photo) standing on a train station platform at night. The subject is posing coolly, holding the strap of an olive backpack with one…
Create a stylish modern travel poster inspired by risography about [City Name]. Use simplified illustrated images of the city: the Gwanganje Bridge, ocean waves, seagulls, beach umbrellas, fish baskets at the market…
Create a stylish modern travel poster inspired by risography about [City Name]. Use simplified illustrated images of the city: the Gwanganje Bridge, ocean waves, seagulls, beach umbrellas, fish baskets at the market…
Create a vertical cozy home scrapbook poster inspired by the reference images. Combine a beautiful realistic interior photograph on the top half with a hand-drawn illustrated memory journal on the bottom half. Top…
Create a sophisticated minimalist editorial art poster featuring [Subject] as the central subject, inspired by premium travel photography and refined hand-drawn illustration. Preserve the subject’s recognizable…
Create a premium 4:5 vertical Instagram travel poster for [Country], designed as part of a cohesive collectible “Souvenir Box” series. Show an open vintage cardboard travel box containing exactly 10 culturally authentic…
Create a stylish illustrated digital collage / character mood-board of a fashionable young African man in a playful sketchbook aesthetic. The composition should feature multiple poses and expressions of the same man…
A cutting-edge commercial 3D isometric illustration of a modern creative workspace, designed for a premium SaaS website hero section. Behance and Dribbble trend-setting style. The composition features a stylized…
Examples are curated from meigen.ai and credited to their original creators.
Aspect ratio and output
Pick the ratio for where the image will live (1:1 stickers and avatars, 4:5 feeds, 9:16 stories, 16:9 banners), then use these five parts to steer the blend.
Say what must not change: "keep the dog's face and red collar".
Framing and angle: centered, full body, flat lay, low angle.
Match the scene or override it: soft studio light, golden hour, neon.
Name the medium if you have no style image: die-cut sticker, enamel pin, felt plush, risograph print.
Exact words in quotes and where they go, e.g. title "SUMMER" at the top.
Keep the subject from image 1 (my corgi, red collar), place it on the beach from image 2, and render it as a glossy die-cut sticker with a thick white border, like image 3. Square, centered, plain background.
FAQ
No. WhiskAI is an independent Whisk-style remixer and is not affiliated with, endorsed or sponsored by Google. Whisk and Google are trademarks of Google LLC. The remix runs on Nano Banana and GPT Image models through licensed API providers.
One remix is one image generation: from 6 credits, depending on model and resolution. The studio shows the exact cost before you generate, and failed images are refunded automatically.
No. Fill at least one slot and you can generate with an empty prompt; the model combines the subject, scene and style by role. A short prompt helps when you want a specific format or text.
Three labelled slots (Subject, Scene, Style), plus extra references on models that accept more, up to 16 in total. GPT Image 2.6 Economy is text-only and has no slots.
Nano Banana Pro is the default: it keeps faces, pets and products faithful and handles text well. GPT Image 2 and 2.5 give bolder, more illustrated looks. Cheaper Nano Banana versions are good for quick drafts.
1K, 2K and up to 4K, depending on the model, in square, portrait and landscape ratios such as 1:1, 4:5, 9:16 and 16:9. Free accounts render at 1K; credit packs and plans unlock higher resolutions.
Yes with any credit pack and with Pro and higher plans, under our Terms of Service, as long as you have the rights to the images you upload. Free and Starter credits are for personal use.
Yes. Press Animate this on any result: the video studio opens in Frames mode with your image as the first frame, and you only describe the motion.
Drop a subject, a scene and a style, or just type. Images start at 6 credits.