AI WORLD HQHow to Generate Images with Google Gemini: A Beginner's Guide
Create beautiful, custom visuals for work or play in minutes using Google's free conversational AI.
Imagine being able to paint a picture just by describing it out loud. Whether you need a quick illustration for a work presentation or a fun graphic for a party invitation, Google Gemini lets you turn your words into custom images in a matter of seconds. This guide will walk you through the simple steps to create your own unique visuals.
- A Google Account: You will need to be signed into a standard Google account. If you use Gmail or YouTube, you already have one.
- Access to Gemini: Open your web browser and go to
gemini.google.com, or open the Gemini app on your mobile device. - Device: This guide works on any computer, tablet, or smartphone with internet access.
- Cost: Image generation in Gemini is currently available for free, though usage limits may apply or change over time with AI tools.
- Time: This guide will take approximately 5-7 minutes to follow.
Open Gemini and start a new chat
First, make sure you're ready to talk to Gemini. When you open gemini.google.com in your web browser, you'll see the main Gemini interface, which typically features a large text box at the bottom of the screen and a "New chat" button somewhere near the top. This is your blank canvas for a fresh conversation.
When you click or tap the "New chat" button, the text box at the bottom will clear, indicating Gemini is ready for your first instruction. This prepares the AI for a new request, ensuring it doesn't get confused with previous topics.
If you don't see a "New chat" button, look for a plus sign (+) icon or a menu option like "Start new conversation" in the top-left corner.
You'll know it worked when the text input area at the bottom of the screen is empty and ready for you to type.
Tell Gemini to generate an image
To create an image, you need to write a prompt (the written instruction you give to an AI, acting like a friendly request to a helpful assistant). Gemini needs to know right away that you want a picture, not just text. Always start your prompt with a clear command for image creation. Type your command into the text box at the bottom of the screen.
After typing your command, press the "Enter" key on your keyboard or click the send icon (often a paper aeroplane or arrow button) next to the text box. Gemini will then process your request, and you'll typically see a "Generating..." message as it works.
If you accidentally just type a description without the command, Gemini might try to answer you with text. Just clarify in your next message, "Please make an image of that."
You'll know it worked when Gemini starts to show a message like "Generating image(s)..." or "I'm working on that for you."
Describe your subject with details
An AI image generator (a tool that turns written text into digital pictures) needs details to work its magic. Think of it like describing a scene to a friend who has their eyes closed. After your initial command from Step 2, continue to type in the main text box, adding specific information about what you want to see.
When you add more detail to your prompt and press Enter or the send button, Gemini will process this enriched request. It will use these clues to build a more accurate and vibrant picture, displaying the results directly in the chat window.
If the images don't quite match your vision, it often means your description was a little too vague. Try adding more adjectives or focusing on specific objects, colours, or actions.
You'll know it worked when Gemini presents you with usually two to four image options that directly reflect the details you provided in your prompt.
Choose your visual style
Gemini can create images in many different artistic styles. If you don't specify a style, it will usually default to a realistic photo. To change the mood or appearance, add style keywords to your prompt in the same text box where you've been typing. This instructs Gemini on the artistic direction for the image.
When you include a style in your prompt and submit it, Gemini will generate new images that incorporate that specific artistic look. You'll see the images appear within the chat interface, often in a grid or gallery format. This allows you to easily compare different styles or refinements.
If the style doesn't look quite right, try using a slightly different keyword (e.g., "cartoon" instead of "3D cartoon") or being more specific (e.g., "impressionist oil painting").
You'll know it worked when the new images Gemini presents clearly show the artistic style you requested, such as brushstrokes for a "watercolour painting" or distinct lines for "line art."
Refine your generated images
Once Gemini generates your images, you might want to make some tweaks. You don't need to start from scratch. You can simply talk to Gemini like a person, typing a new instruction in the text box that refers to the images it just made. This tells Gemini to adjust the previous output rather than starting a completely new image set.
After you send your refinement request, Gemini will consider your feedback and generate updated versions of the images, incorporating the changes you asked for. These new images will appear in the chat, usually below your request.
If Gemini seems confused by your request, try being more specific about which image you want to change ("the last one," "the first image") or what exact element needs altering ("the background," "the animal").
You'll know it worked when Gemini displays new images that show the specific adjustments you asked for, like a bluer sky or a newly added sailboat.
Save or share your image
Once you have an image you like, you'll want to save it or share it. Look for the image within the chat window. Usually, if you click or tap on an image, it will open in a larger view, where you'll find options to download or share.
When you select the download option (often an arrow pointing down, or a "Download" button), the image will be saved to your device's downloads folder or photo gallery. If you choose to share, a menu might appear allowing you to copy the image or share it directly to other applications.
If you don't see a clear download button, try right-clicking (on a computer) or long-pressing (on a mobile device) the image itself; this usually brings up a context menu with "Save image as..." or "Download image" options.
You'll know it worked when the image file appears in your device's downloads or photo library, ready for you to use or share.
- Being too brief: Writing "a cat" will give you a random result. Instead, give Gemini specific details.
- The fix: Writing "a sleepy tabby cat curled up on a knitted blue blanket in front of a warm fireplace" gives Gemini the clues it needs to make something beautiful and specific to your idea.
- Asking for readable text: AI image tools sometimes struggle to spell words correctly or make them look natural inside pictures.
- The fix: It is usually best to generate the image without any text, and then add your words later using a simple design app or photo editor.
- Using negative words: AI models respond better to what should be in the picture, not what shouldn't.
- The fix: Instead of saying "a kitchen with no clutter," try saying "a clean, tidy, minimalist kitchen with polished benchtops." Always frame your request in a positive, descriptive way.
Let’s make your very first image. Copy and paste the prompt below into Google Gemini right now to see what happens:
"Create a simple, cheerful cartoon illustration of a happy koala holding a warm cup of tea."
✦ Original step-by-step guide by AI World Co.'s AI editorial team. Written in plain language, reviewed for accuracy.
← Back to all stories