Google Gemini can create and edit images, and you do not need any special skills to start. This guide shows where to find image generation, how creating differs from editing, what to know about limits and quality, and ten starter prompts you can copy today.
Quick answer
To make an image in Gemini, open the app or gemini.google.com, start a new chat, choose image creation, describe the picture in plain sentences and send. To edit your own picture, upload it first and say what to change and what to keep. Limits and models change, so check the app.

Where to find image generation in Gemini
Gemini lives in a web app at gemini.google.com and in mobile apps. Image creation is part of the same chat.
- Open Gemini and sign in with a Google account.
- Start a new chat.
- Look for an option to create images, or simply describe an image and ask Gemini to create it.
- If a model picker is shown, choose the image model offered. People often call it Nano Banana. Names change, so check what your app shows.
Menus move often. If you cannot find an option, ask Gemini directly: "Create an image of a quiet lake at sunrise".

Create vs edit
Create means you describe a new picture from words only. This is good for places, objects, art styles and characters that are not you.
Edit means you upload a photo and ask for a change. This is good for new backgrounds, outfits, lighting or a different style. When you edit, always say what must stay the same, for example: keep my face, skin tone, hair and body proportions exactly the same as in the photo I upload.
A quick rule: if the subject is your real face, use edit and an identity-lock line. If it is an imaginary scene, use create.
Your first image in five steps
- Open a new chat.
- Type a short description with subject, setting, light and mood.
- Send it and wait for the image.
- Ask for one change at a time, such as "same picture, golden hour light".
- Download the version you like.
Short prompts work for a first try. Add detail when you want control.
Limits and quality settings
Gemini has usage limits, image size limits and feature differences between free and paid plans. These change over time and vary by region, so we do not list numbers here. Check the Gemini app and Google's help pages for what applies to you today.
What you can control yourself:
- Detail. Name the light, the lens and the setting.
- Framing. Say portrait, square or wide, if the app lets you choose.
- Quality. Start from a sharp photo when editing.
- Retries. Ask for variations if the first result is close.

10 starter prompts
Copy any of these. Replace the details with your own.
- Create an image of a small round table with a white teapot and two cups in a sunny greenhouse full of ferns, soft morning mist, realistic photograph.
- Create an image of a red bicycle leaning against a blue door in a narrow street, overcast daylight, 35mm lens.
- Create an image of a bowl of peaches on a wooden table by a window, soft light, macro lens, fresh and quiet.
- Create an image of a paper cut forest with layered green shadows and a small fox in the middle.
- Create a watercolour painting of a quiet lake with pine trees and low morning mist.
- Create a studio portrait of a friendly woman in a cream blazer against a plain grey backdrop, soft key light, 85mm lens. (Use a photo prompt from our library if you want your own face.)
- Create a 3D render of a small robot gardener holding a watering can, soft studio light, playful.
- Create a cozy cafe interior with warm lamps, a wooden counter and rain on the window, evening, wide shot.
- Create a cinematic street scene at dusk with wet roads and warm shop lights, shallow depth of field.
- Create a hand drawn anime illustration of a girl with a paper lantern at a night market.
For photo edits, browse the Gemini photo prompts and filter by style.
3D drawings and stylised looks
Many people ask how to make a 3D drawing in Gemini. Describe the look in plain words: a smooth 3D render with soft studio light and rounded shapes, like a collectible toy. Our figurine prompts show full examples.
Tips that make results better
- Be specific. Colour, fabric and place beat adjectives like "beautiful".
- One change at a time. Keep what works.
- Keep it safe. Avoid real people, brands and sensitive topics, which can cause refusals.
- Use a sharp photo for edits. Front-facing, in daylight.
- Save prompts you like. Use the bookmark on our site.
Common problems
Gemini answers in words. Reply "create the image now".
The face changed. Send the identity-lock line alone.
The image has odd text. Skip text, or keep it short and quoted.
A refusal appears. Rewrite in neutral words.
Find more fixes in our help article.
Where to go next
When you are comfortable, try the Gemini Prompt Generator to build photo-editing prompts from a few choices, or read how the six-part prompt formula works. If you want a prompt written for you, use the free request form.
A note on safety and sharing
Only upload photos of people who agreed. Do not make images that could mislead people about something real. When you share an AI image, say it is AI generated if there is any chance people could take it as real. Check your result for mistakes before you post it, especially hands, text and small objects. If you plan to use an image commercially, read the current terms for the tool and for any model it uses, because rules can differ by plan.
A first week practice plan
On day one, make three images from words only. On day two, edit one photo of yourself with a background change. On day three, try an outfit change. On day four, test a lighting change. On day five, try a style change such as vintage film. On day six, repeat your favourite with a different prompt version. On day seven, save your best prompts and write down what worked. By the end of the week you will know how Gemini responds to your style of writing.
Editing starter prompts (with your own photo)
Upload one clear photo, paste one of these and add the identity-lock line.
- Change my background to a quiet beach at sunset and keep everything else the same.
- Put me in a tailored navy blazer over a white shirt against a plain grey backdrop, soft window light.
- Turn my photo into a vintage 35mm film picture with warm faded colours and fine grain.
- Make a cinematic version of my photo with a subtle teal and orange grade and shallow depth of field.
- Turn me into a collectible 3D figurine on a desk with a blank box beside it.
How to describe a scene so Gemini understands
Imagine you are briefing a photographer who cannot see your mind. Say who is in the picture, what they wear, where they are, what the light is doing and which lens you want. Use plain words. If a detail matters, say it once, clearly. If something must not change, say so directly. Gemini follows short, concrete sentences better than long, flowery ones.
Common beginner questions
Do I have to write long prompts? No. A short prompt gives a decent first result. Add detail only where you want control.
Why does the same prompt give different images? Image models include some randomness. Run it again, or ask for a variation.
Can I reuse a prompt? Yes. Save the ones that work and change only the details.
What if I do not like the result? Say what to change in one sentence, such as "same picture, brighter background". Gemini keeps the rest.
Should I use one long chat? Use one chat per idea. A fresh chat avoids confusion from earlier messages.
Sizes and shapes in plain words
Gemini and other tools may let you choose the shape of the image. A tall shape suits portraits and phone screens. A square shape suits profile pictures. A wide shape suits landscapes and video thumbnails. If the app does not offer a choice, mention the shape in your prompt, for example "a tall portrait shot".