To write an image prompt in Breme, open Create inside your project folder, choose Image, and expand the prompt editor. Describe the subject and action, add the setting, then specify framing and one clear lighting treatment. Choose Done, set the model, aspect ratio, and resolution in their controls, review the credit cost, and generate. Check the result against your brief and refine one missed instruction at a time with the same model and settings.
Before you start
- A Breme project folder
- Sufficient credits and access to an image model
Open the expanded image prompt editor
Inside your project folder on Project canvas, open Create and choose Image. Choose Expand prompt so you can read the brief as it grows. The recorded example starts from text without an attached image.
Name the subject and action
Start with one black-and-white dog running through shallow water, with splashes around its paws. Naming the subject and a specific moment gives you observable details to check in the result.
One black-and-white dog running through shallow water, splashes around its paws.
Add the setting and palette
Keep the subject and action, then add the dark teal forest reflected in the water. This defines the surroundings and background palette. Add each new instruction to the existing brief.
A dark teal forest reflects in the water.
Describe framing and space
Add a full-body side view at the dog’s eye level. Place the dog on the left with open space ahead on the right, and ask to keep all four paws in frame. This gives the action room and makes the pose easier to inspect.
Full-body side view at the dog’s eye level, with the dog on the left and open space ahead on the right. Keep all four paws inside the frame.
Choose lighting and useful detail
Add soft overcast daylight, photorealistic wet fur, and crisp water droplets. One clear lighting treatment keeps the instructions consistent, while concrete surface details describe the finish you want. The complete brief below combines all four stages.
One black-and-white dog running through shallow water, splashes around its paws. A dark teal forest reflects in the water. Full-body side view at the dog’s eye level, with the dog on the left and open space ahead on the right. Keep all four paws inside the frame. Soft overcast daylight. Photorealistic wet fur and crisp water droplets.
Set the model and image controls
Choose Done to return to the settings. Choose Seedream 5 Pro, set Aspect ratio to 16:9, and leave Resolution at 2K, as shown in the recording. The wide frame gives the running dog room, and this resolution is the example’s detail-review setting. Review the displayed credit cost, then press Generate. The tutorial trims the generation wait.
Keep the model and settings fixed when comparing prompt revisions. Choose technical settings in their controls rather than relying on prompt text.
Inspect the result against the brief
Open the completed image in the app. Check the dog, paws, splashes, forest background, and space ahead. The prompt guides the image and does not guarantee every detail. If one part misses the brief, revise that instruction and compare another version using the same model and settings.
Review image export formats
Close the preview, select the result, and choose Edit in its item toolbar. Choose Export in the panel on the left to review formats. The recorded result is 2816 × 1584 and offers PNG, JPG, and WEBP. Choose the format you need when you are ready to save. Reuse the subject, action, setting, framing, and lighting questions for your next scene.
Common questions
Tools in this guide: AI Video Generator, AI Image Generator


