The Gemini photo generator is the image feature inside the Gemini app, and it runs on Google’s Nano Banana models. You type a description or upload a photo, and Gemini draws a new image or edits yours. It is quick and free to try, but it has real limits: daily allowances, rules about people, and a few things it simply won’t do. This guide covers what the Gemini photo generator can and can’t do, so you know what to expect before you spend an afternoon on it.
Gemini photo generator: the short answer
Gemini can create images from text and edit photos you upload, and every image carries Google’s invisible SynthID watermark. Free accounts get the faster Nano Banana 2 model within a daily or time-based allowance. Paid Google AI plans raise the allowance and add the heavier Nano Banana Pro model for hard jobs. It will refuse some requests, mainly anything that targets real, named people or breaks Google’s content rules.
Key takeaways
- You can generate images from a text prompt or edit a photo you upload.
- The default model for most people is Nano Banana 2. Nano Banana Pro is a paid extra.
- Limits depend on your plan, and Google now describes them in terms of a shared usage meter rather than a fixed number.
- Gemini won’t make images of named real people from a text prompt, and it blocks sexual, violent or child-related content.
- Good prompts are specific about subject, setting and style. Edits work best when you also say what must stay the same.
What the Gemini photo generator can do
Most people discover it by accident. You ask Gemini for “a picture of a lighthouse at dusk” and an image appears in the chat. That is the simplest use, but it does more than that.
| Task | What it looks like in practice |
|---|---|
| Text to image | Describe a scene and get an original picture |
| Photo editing | Upload a photo, then change the background, clothing, colours or lighting |
| Combining images | Merge several uploaded pictures into one scene |
| Style changes | Turn a photo into a watercolour, sketch, toy or film still |
| Text in images | Add short words to posters, signs and mock-ups |
| Follow-up edits | Keep chatting to adjust the result without starting again |
That last row matters more than it looks. Gemini remembers the conversation, so “make the sky more orange” or “remove the person on the left” works as a small correction on the image it just made. Our Nano Banana prompt guide explains how to word those requests so the model changes only what you asked.
How to use the Gemini photo generator, step by step
- Sign in. Open the Gemini app or the Gemini website with a Google account. Image creation needs you to be signed in.
- Choose the image tool. Use the image option in the prompt area, or just write “create an image of…” in the chat.
- Write a clear prompt. Give the subject, the setting and the style in a sentence or two.
- Upload a photo if you’re editing. Add the picture, then describe the change and what to leave alone.
- Review and refine. Ask for one adjustment at a time. Download the version you like.
Menu names change from time to time, so don’t worry if the buttons look a little different on your phone. The flow stays the same.
Which model do you get: Nano Banana 2 or Nano Banana Pro?
Most free users get Nano Banana 2, which is built on Gemini 3.1 Flash Image and was announced on 26 February 2026. It’s fast and good enough for everyday pictures. Nano Banana Pro, built on Gemini 3 Pro Image, is the slower, more careful model. People use it for crowded scenes, accurate text and detailed edits. We compare both in our Nano Banana Pro guide.
Here’s a rule of thumb that saves allowance. Start every idea on the default model. Only switch to Pro when a prompt keeps failing in the same way, for example the text is misspelled three times in a row.
How many images can you make?
This is the part that confuses people, because the numbers have changed. Until May 2026, Google’s help page printed daily counts. A third-party tracker that kept copies of that page lists these last published figures:
| Plan | Nano Banana 2 | Nano Banana Pro |
|---|---|---|
| Free | Up to 20 images a day | Not included |
| Google AI Plus | Up to 50 a day | Up to 50 redos a day |
| Google AI Pro | Up to 100 a day | Up to 100 redos a day |
| Google AI Ultra | Up to 1,000 a day | Up to 1,000 redos a day |
Since then, Google has described image use as part of a shared usage meter that refreshes about every five hours, up to a weekly limit, and the page no longer lists a number. Paid plans get a bigger share of that meter. So treat the table as a rough guide to the old daily caps, and trust the notice inside your own app, which shows when your allowance refreshes. Two small details help: reports say failed requests aren’t counted, and a request can return fewer images than you asked for.
What the Gemini photo generator can’t do
Knowing the limits upfront saves frustration. Here are the common walls people hit.
- Named real people. Ask for a picture of a specific politician or celebrity from text alone and Gemini will usually refuse. Google’s rules also ban using someone’s image without consent or impersonating a person to deceive.
- Sexual, violent or child-related content. These are blocked.
- Brand logos and copyrighted characters. Text-only requests often get declined, and you remain responsible for what you publish.
- Perfect text in every case. It’s far better than early image models, but long sentences can still come out wrong.
- Exact control over small details. Hands, crowds and complex machinery can still look off.
- Passport or ID photos. An AI-edited picture is not a legal identity photo.
Age and location rules also apply. Third-party summaries of Google’s policy say image creation needs a signed-in account aged 13 or over, and that editing an uploaded photo needs you to be 18. Work and school accounts may have extra limits set by an administrator. Check Google’s current help page if it matters for your account.
Why Gemini refuses a harmless prompt
Sometimes a perfectly ordinary request fails. A few things usually explain it.
- You’ve hit your allowance. Try a plain prompt such as “a bowl of fruit”. If that fails too, it’s a limit or an outage, not your wording.
- A name slipped in. Remove names of real people and describe an invented person by age range, clothes and setting.
- The filter misfired. Safety filters err on the side of caution. Reword the prompt calmly. Don’t submit near-copies of a refused request over and over, because trying to get around a refusal can itself break the rules.
- The account type blocks it. Work, school and under-18 accounts can have image features switched off or limited.
Prompts to try today
Copy these and swap in your own details. They work because each one names a subject, a setting and a look.
- Product shot: “A white ceramic mug on a wooden table by a window, soft morning light, shallow depth of field, clean magazine style.”
- Edit your photo: “Change the background to a plain light grey studio wall. Keep the person, pose and clothing exactly the same.”
- Illustration: “A cosy reading nook with a green armchair and a cat, flat vector illustration, warm colours.”
- Social post: “A bright poster that says ‘Open Mic Night’ in bold hand-drawn letters, purple and yellow, plenty of space around the text.”
- Restyle: “Turn this photo into a soft watercolour painting, keeping the composition the same.”
For a larger bank of ideas, see our 50 Gemini prompts, and browse more picture guides in the Gemini and Nano Banana section.
Are Gemini images safe to use?
Every image made in the Gemini app carries SynthID, an invisible watermark that lets Google tools detect AI-generated content. That’s good for trust, but it doesn’t make an image free of risk. You are still responsible for what you publish. Don’t use generated pictures to mislead people, copy another creator’s work or imitate a real person. For commercial use, read Google’s current terms and keep your own notes on how the image was made. Our disclaimer explains that this site is independent and not affiliated with Google.
Frequently asked questions
What is the Gemini photo generator?
It is the image feature inside the Gemini app. You describe a picture or upload a photo, and Gemini creates a new image or edits yours. It runs on Google’s Nano Banana models, and every image carries an invisible SynthID watermark that helps tools detect AI-made content.
Is the Gemini photo generator free?
Yes, there is a free tier that uses Nano Banana 2 within a usage allowance. Older Google help pages listed up to 20 free images a day, but Google now describes limits as a shared usage meter. Paid Google AI plans raise the allowance and add Nano Banana Pro.
Can Gemini edit my own photos?
Yes. Upload a picture, say what to change and what to keep, and Gemini will edit it. Examples include swapping a background, changing clothing or restyling the picture. Summaries of Google’s policy say editing uploaded photos needs an adult account, so check the current terms.
Why won’t Gemini generate images of people?
Common reasons are a request for a named real person, a rule violation, a hit usage limit, or an account type with image features limited. Test a plain prompt such as a bowl of fruit. If that works, remove names and describe an invented person instead.
How many images can I make with Gemini per day?
Google’s last published numbers, seen in May 2026, were 20 a day on the free plan, 50 on AI Plus, 100 on AI Pro and up to 1,000 on Ultra. Google has since moved to a shared usage meter without a printed count, so check the notice in your app.
Can I use Gemini images commercially?
Often yes, but you are responsible for what you publish. Do not copy brands, characters or real people, and read Google’s current terms before commercial use. Keep a record of your prompts, and add clear labels where honesty about AI use matters.
Is the Gemini photo generator worth using?
Yes, if you want fast concept images, simple edits and social graphics without learning design software. It’s less suited to work that needs exact control or strict legal clearance. Try a few prompts on the free tier, learn where it struggles, and only then decide whether a paid plan helps. You can see how we judge tools like this on our how we test page, or head back to the Google Edits homepage for more guides.
Independent publication. Not affiliated with or endorsed by Google. Limits and prices change; verify in your own account.
