
How to Use Google's Nano Banana: A Beginner's Guide
A plain-language walkthrough on how to use Google's Nano Banana in the Gemini app, from your first prompt to editing photos and keeping faces consistent.
Google's Nano Banana is the image model living inside the Gemini app, and despite the goofy name it's genuinely one of my favorite tools right now. If you're wondering how to use Google's Nano Banana without getting lost, good news: it's mostly just typing what you want in plain English. Let me walk you through getting started, writing prompts that work, and the editing tricks that make it feel a little magic.
I'll keep this practical. By the end you'll know how to make your first image, edit a real photo, keep a person looking like themselves across versions, and blend multiple pictures together. No technical background needed.
Step One: Open It in the Gemini App
Inside Gemini, open the tools menu and pick the "Create images" option. From the model menu you can choose Fast, Thinking, or Pro depending on how much quality and care you want. Fast is great for quick experiments, and Pro is where you go when the result really matters.
Then you do one of two things: type a prompt to make something new, or upload an image you want to edit. That's the whole on-ramp. No setup, no settings to wrestle with first.
Step Two: Write Your First Prompt
The starting formula Google suggests is simple and it works: subject, action, scene. Something like "Create an image of a cat napping in a sunbeam on a windowsill." Type that, generate, and you've got your first image.
From there, the single best habit is adding detail. Instead of "a woman in a red dress," try "a young woman in a red dress running through a park." The more specifics you give, the better Nano Banana follows you. Think about:
- Subject and action: who or what, and what they're doing.
- Scene and location: where it's happening.
- Style and composition: photo, illustration, close-up, wide shot.
Step Three: Edit Photos by Just Asking
This is the part that hooked me. Once you have an image, or after you upload one, you can edit it through conversation. You literally tell Gemini what to change. Knowing how to use Google's Nano Banana for edits is mostly about being specific with your requests:
- "Change the background to a snowy mountain."
- "Replace the coffee cup with a glass of orange juice."
- "Add a pair of sunglasses."
The clever bit is that it alters the part you mention while preserving the rest. So you can adjust one element and keep everything you already loved about the picture. If you don't like the result, just ask again with a tweak. It's a back-and-forth, not a one-shot gamble.
Step Four: Keep a Person or Pet Consistent
One of Nano Banana's standout tricks is keeping a face consistent across edits. You can take the same person or pet and put them in new scenes, outfits, or styles, and they still look like themselves. That's surprisingly hard for image AI, and it's a big reason this tool stands out.
To use it, generate or upload your subject, then ask for changes around them: a new background, different lighting, a fresh setting. Because it holds onto identity, you can build a whole series that feels like the same character throughout. It's great for consistent profile images or telling a little visual story.
Step Five: Blend and Style-Match Multiple Images
You can also combine pictures. Nano Banana lets you blend multiple photos together, or take the style from one image and apply it to the subject of another, all while keeping the important details intact. The Pro version goes further and lets you mix up to 14 reference images in a single prompt, which is wild for complex composites.
A practical example: upload a product photo and a background you like, then ask Gemini to place the product into that scene. Or hand it a photo plus a painting and ask it to render your photo in that painting's style. Describe what you want clearly and let it do the merging.
A Few Prompt Tricks That Help
Once you're comfortable, these small additions consistently improve my results:
- Describe the lighting. Asking for a "three-point softbox setup" gives product shots that even, studio-lit look.
- Name camera angles. "Low angle" or "overhead" changes the whole feel of a shot fast.
- Be explicit about text. If you need words in the image, spell out exactly what they should say, since Pro handles text well.
Picking Fast, Thinking, or Pro
The model menu trips up a lot of beginners, so here's how I actually choose. Each option is a different balance of speed and quality, and you'll move between them depending on the moment, not pick one forever.
- Fast: my go-to while I'm still figuring out a prompt. Quick results mean I can iterate ten times in the span it'd take Pro to finish once.
- Thinking: a middle ground when I want more care than Fast but don't need the full Pro treatment.
- Pro: what I switch to for the final render, anything with text in it, or complex composites with lots of reference images.
The workflow that works for me: draft on Fast, nail the prompt, then run the keeper on Pro. You get the best of both, quick iteration and a polished final image, without waiting around for every experiment.
Common Beginner Mistakes to Dodge
A few things I wish someone had told me on day one:
- Vague prompts. "A nice view" gives you something generic. Name the place, the time of day, the mood, the style.
- Starting over instead of editing. If a result is 90% right, ask for a small change rather than regenerating from scratch. The conversational editing is the whole point.
- Forgetting the consistency trick. If you want the same character across images, tell Gemini to keep them consistent rather than describing them fresh each time and hoping.
- Burning your free tries on tests. The free app only gives three low-res generations, so save them for a real attempt once you know roughly what you want.
Who Should Skip It
I'll be straight with you. If you only ever need one quick image, the free app's three low-res tries may be all you touch, and that's fine. And if you want frame-perfect manual control over every pixel, a real editor will serve you better than any prompt-based tool. Nano Banana shines at fast, conversational creation and editing. It's less suited to work where you need to nudge things by hand down to the exact detail.
And like every image model right now, it can still flub small text, hands, or a cluttered scene now and then. That's not a flaw unique to this tool, it's just where the technology is. The good news is that fixing it is usually one more sentence: ask for the change, and let it try again. Going in with that expectation keeps the experience light instead of leaving you annoyed at a stray sixth finger.
The Bottom Line
Learning how to use Google's Nano Banana comes down to a friendly loop: open Create images in Gemini, write a specific prompt, then edit by simply asking for changes. Lean on its consistency trick to keep faces looking right, blend photos when you need a composite, and add lighting or angle details to level up your prompts. Start with the Fast model and plain descriptions, get comfortable, then reach for Pro when quality counts. It's one of the most beginner-friendly image tools I've used, and you'll be making things you actually like within the first ten minutes.
Emily in AI
Emily in AI is a plain-English guide to AI tools, tips, and beginner guides. Every tool gets tested and written up without the hype or the jargon, so you can figure out what actually helps. New posts every week.
About Emily in AI →

