The page could not load. Please refresh to try again.
This is taking longer than expected. You can keep waiting or refresh the page.
TEXT TO IMAGE
Describe a scene, a character, or an idea. Choose an image model in MagicCanvas and turn your description into a picture you can refine.
Text to Image creates new pictures from a description. Use it to explore a composition, test an atmosphere, or develop a concept before refining the details.
Explore lighting, color, and composition for a campaign before planning a shoot. Treat generated product details as concepts, and verify them before using an image to represent a real product.
A sculptural violet perfume bottle on glossy black volcanic stone, an arc of water, lavender rim light and warm sunlight, realistic glass reflections, luxury campaign photography.Try this prompt →
Build a moodboard for a story, presentation, or travel-inspired design. Describe the atmosphere and viewpoint. Generated scenery is fictional, even when it resembles a real location.
A futuristic bronze observatory above a sea of clouds, lush gardens, a peach-colored moon on the horizon, indigo twilight, warm windows, cinematic environment concept art.Try this prompt →
Try a character idea for social posts or a visual story. Specify the character’s appearance and activity. For a series, review each result for changes in shape, color, and facial details.
A tiny orange fox astronaut in a cream spacesuit and bubble helmet, tending a glowing plant on a lavender moon, tactile fur, curious eyes, warm cinematic rim light, polished 3D illustration.Try this prompt →
A style word gives the model a direction. Add a clear subject, a setting, and a few visual details to make the request more specific.
A ceramic coffee cup on a wooden table, window light, shallow depth of field, realistic editorial photography.
Try photography →
A small seaside bookshop with plants by the door, warm pastel colors, hand-drawn editorial illustration.
Try illustration →
A tiny floating garden inside a glass sphere, soft studio lighting, playful 3D render, lavender background.
Try 3d render →
A mountain village after rain, mist between the houses, muted blue and green watercolor washes on textured paper.
Try watercolor →These examples add style instructions to your prompt. Available generation controls depend on the image model you choose.
HOW IT WORKS
Start with the main subject and setting. Add the viewpoint, lighting, and style you want. You can write a prompt above or try one of the examples.
Continue to Create and pick an available Text to Image model. Choose supported output settings and a shape that fits your project. Review the generation cost.
Look at the composition first, then check smaller details. Change one part of the prompt at a time so you can see what helped. Results can vary between generations.
Download the version you want. Use an image-editing tool for a targeted change, or bring the finished image into Image to Video to explore motion.
Compare output sizes and aspect ratios before choosing a model. For a social post, look for 1:1 or 4:5; for a banner, look for 16:9 or wider. Model availability and settings can change; Create shows the options and price available to your account.
Try this structure: subject + setting + composition + light + style. Start with the part of the image that matters most. Avoid asking for several conflicting styles or unrelated subjects in the same request.
A coffee shop.
A small corner coffee shop on a rainy street, warm light through the windows, eye-level view, muted colors, watercolor illustration.
Leave space for a headline when planning a banner. For precise lettering, add the final text in a design editor. Check faces, hands, product labels, and repeated patterns before publishing a generated image.
Use a focused editing tool when you want to change one part of a result. Keep your original image so you can compare versions as you work.
It creates a new image from a written description. Your prompt guides the subject, setting, visual style, and composition. It generates a result rather than searching for a matching stock photo.
Describe your idea on this page and continue to Create. Select an available image model and settings, review the cost, then generate. Inspect the result and refine the prompt if you want another version.
No. Text to Image starts from a description. If you already have a photo you want to change, use Image Edit in Create or AI Replace instead.
The model section lists the Text to Image options returned by the current catalog. Availability, output sizes, and costs can change or depend on your account. The model picker in Create shows the options available when you generate.
Generation uses credits or diamonds. The amount depends on the model and settings. Check the price next to Generate in Create and review the plans page for account options.
Describe a style in your prompt, such as photography, watercolor, or 3D illustration. Choose an aspect ratio from the settings supported by your selected model in Create. Style examples on this page add prompt text; they are not separate model presets.
Text quality varies by model and prompt. Check spelling and letter shapes carefully. For exact headlines, labels, or legal text, leave space in the image and add the final typography in a design editor.
Not necessarily. Generations can vary even with similar instructions. Keep a copy of prompts you like and compare results before deciding which version to use.
Check MagicCanvas’s terms, the selected model’s applicable terms, and any third-party rights relevant to your intended use. Do not assume that generated logos, likenesses, or characters are cleared for commercial use.
Yes. Download or select your finished image and use Image to Video to describe how it should move. Review the video for changes to the subject and other details.
Inspect hands, faces, text, product details, and repeated patterns. AI images can contain errors or resemble existing work. Make sure the image fits your intended use and avoid presenting a generated scene as documentary evidence.
Start with a few words. Refine the image as your idea takes shape.
Create an image →