Short answer
Open a chat in ChatGPT or Gemini and describe what you want: the subject, the scene, and the style. The AI draws it. You do not need extra software, only an account. In Google's Gemini help page, image generation with Nano Banana is listed for every tier, including people with no paid plan. Free limits change often, so check the app. The easiest way to start is to make one image, then fix the parts you do not like by asking in plain words.
Where you can make images
Other tools exist too, such as Midjourney and Adobe Firefly. This guide sticks to Gemini, which I checked against Google's official pages, and ChatGPT, which is widely used. Menus and screens change, so yours may look a little different.
Google's help page says Gemini usage limits can change, and that image and video requests use more of your allowance than plain chat. So do not memorize a number. If you hit a limit, wait and try later, or start a new chat.
ChatGPT
- How to make an image
- Type what you want to see in the chat, for example "draw a cat sleeping in a sunny window." You can also upload a photo and ask for changes.
- Free use
- Free accounts have been able to generate images, but limits change often. Go by the message the app shows you.
Gemini
- How to make an image
- Describe the image in the chat, or upload a photo and say "turn this into a watercolor." Google calls its image model Nano Banana.
- Free use
- Google's help page lists image generation for all tiers, including no plan. Check your usage limits in the app.
The five-part prompt
A prompt is the message you send to the AI. It does not need to be long or fancy. If you cover these five things, results get much more predictable.
You can state the aspect ratio in words: "square," "wide 16:9," or "vertical 9:16 for Shorts." If the result is the wrong shape, just ask again.
Subject
- What to write
- What to draw
- Example
- One gray tabby cat
Scene
- What to write
- Where it is and what it is doing
- Example
- Curled up asleep in a sunny window
Style
- What to write
- Photo, watercolor, flat illustration, etc.
- Example
- Soft watercolor illustration
Framing
- What to write
- How close, what angle
- Example
- Full body, seen from above
Shape and use
- What to write
- Aspect ratio and where it will be used
- Example
- Square, for a profile picture
Example 1: a cat illustration
You can describe a cat in words without uploading a photo. The more specific you are about fur color and markings, the closer it looks to your cat.
Draw a gray tabby cat with four white paws, curled up asleep in a sunny window. Soft watercolor illustration, warm beige tones, square format, simple background.
Add eye color, nose color and ear shape to make it look more like your own cat.
Example 2: a cafe menu image
This makes a food-photo style image for a menu or an Instagram post. AI often misspells words inside images, so ask for no text and add the text later in another app.
Create a photo-style image of a vanilla latte, the signature drink at Corner Coffee. A white mug on a wooden table, shot from a 45-degree angle, morning sunlight, softly blurred cafe interior in the background. No text. Wide 3:2 format.
The image may not match the drink you actually sell. If you use it in an ad, add a line like "image for illustration only."
Example 3: a birthday card
Add the message yourself with your phone's photo editor or a design app. When AI writes text inside a picture, the letters are often wrong or garbled.
Make a birthday card image for my friend Jordan. Pastel balloons and a small cake, with a large empty space in the middle so I can add my own message later. Vertical 4:5 format, no text.
Saying where you want empty space makes the image easy to edit later.
Editing your own photo
Google says can keep the look of the people and pets in a photo while changing clothes, backgrounds or style, blend several photos into one scene, and make edits over several rounds. ChatGPT can also edit a photo you upload. The steps are the same in both.
- Upload the photo with the attach (+) button in the chat box.
- Say what to change and what to keep.
- Look at the result, then ask for one fix at a time, such as "make only the background brighter" or "keep the ear shape."
- Save any result you like right away. In a long chat, the image can drift away from the original.
Keep the cat's face and fur pattern exactly as in the photo I uploaded. Change only the background to a snowy window view. Make the mood warm and festive.
Be careful with real people's faces
- Do not upload or edit another person's photo without their permission. Sexual or humiliating edits can be illegal.
- Using an image that looks like a celebrity, politician or other well-known person in an ad or promotion can cause legal trouble. Some services refuse these requests.
- An AI image presented as a real photo can mislead people. When you post one, it is safer to say it was made with AI.
- Google says images created or edited in the Gemini app carry a visible watermark and an invisible watermark.
When you do not like the result
- Ask for one change at a time instead of several.
- Be specific. Instead of "make it prettier," say "simplify the background" or "use softer lighting."
- Say what you do not want: "no text," "no people."
- If it has drifted far from what you wanted, starting a new chat is usually faster.
- Fine details like fingers, lettering and clock numbers are often wrong. Zoom in and check.
Taking an image further into video
You can also turn an image into a short video. The video basics guide compares the tools, and the guide on labeling and monetizing AI video covers what to disclose when you post. Browse the make section for step-by-step projects.
Related terms
- Prompt
- The message you send to the AI. For images, it covers the subject, scene, style, framing and aspect ratio.
- Nano Banana
- The name Google uses for the image generation and editing model in the Gemini app.
- SynthID
- An invisible watermark Google adds to images made with its AI.
- Aspect ratio
- The ratio of width to height. Square is 1:1, Instagram portrait is 4:5, Shorts and Reels are 9:16.
FAQ
How many images can I make for free?
It depends on the service and the day, and the limits can change without notice. In Gemini, check the usage limits in the app. In ChatGPT, read the message that appears when you reach a limit. There is no fixed number to rely on.
Can I use AI images to promote my business?
Usually yes, within the service's terms. If an image resembles someone else's character, logo or a real person, it can cause problems, so check the terms of the tool you used and avoid look-alikes.
The text inside my image is garbled. Why?
AI still struggles to spell words correctly inside images. Ask for no text and add the words afterward in a photo editor or design app.
Can I upload photos of my kids?
Some services can use uploads to improve their models unless you change a setting, so check the privacy settings first. It is safest not to upload photos that show a child's face.
Do I need to be good at drawing?
No. You only need to describe what you see. Think of it as explaining a picture to a friend over the phone.
Keep going
Checked against sources on 2026-11-05. Sources: Google, Nano Banana image editing in the Gemini app, Gemini Apps Help: usage limits and plans. How we write