AI Image Prompt Builder
Pick a style, light and camera, see what they look like, and copy a prompt written the way ChatGPT, Gemini or Midjourney want it.
How this format is written
- Written as plain sentences that name the medium, lighting and colors, with the aspect ratio stated in the prompt. OpenAI docs
- What to leave out is listed at the end as “No …”. OpenAI docs
Style
10 looksLighting
how the scene is litShot
how much is in frameCamera angle
where the camera isLens
for photosColors
the overall paletteShape
aspect ratioLeave out
optionalCreate a photorealistic photograph of [your subject]. Medium shot from an eye-level angle. Warm golden-hour sunlight. Aspect ratio 3:2 (landscape).
How the prompt is put together
Each choice adds a short, plain description to the prompt: the medium, then the shot and angle, the light, the colors and your details. How those pieces are arranged depends on the generator, and follows its own documentation:
- ChatGPT: Written as plain sentences that name the medium, lighting and colors, with the aspect ratio stated in the prompt. What to leave out is listed at the end as “No …”.
- Gemini: Follows Google’s template: the shot and subject, then the light, then “Shot from a [camera angle] with a [lens type]”, then the aspect ratio. Google advises describing the scene you want instead of writing “no …”, so the things to avoid aren’t added. Describe the alternative in Details instead (“an empty street” rather than “no cars”).
- Midjourney: The aspect ratio is added at the end as --ar width:height, in whole numbers. Things to leave out go in the --no parameter, as a comma-separated list. Parameters go after the prompt text, with nothing written after them.
We re-read these pages every day and mark a format here if its page has changed. Sources:
- OpenAI: GPT Image prompting guide
- Google: Gemini image generation
- Midjourney: Aspect Ratio
- Midjourney: No
- Midjourney: Parameter List
About the preview
The picture is a simple drawing that redraws itself for each choice, so you can see what “low angle” or “backlit” means before you use the words. No AI made it, and a generator’s result will look different.
What the words mean
Shots
- Extreme close-up: one detail fills the frame, such as an eye or a hand.
- Close-up: a face, or head and shoulders.
- Medium shot: from about the waist up.
- Full body: the whole subject, head to feet.
- Wide shot: the subject small within its surroundings.
Camera angles
- Eye level: the camera at the subject’s eye height; feels neutral.
- Low angle: looking up at the subject, which makes it feel bigger or more powerful.
- High angle: looking down, which makes the subject feel smaller.
- Bird’s-eye view: straight down from above.
Lenses
A wide-angle lens takes in more of the scene. Longer lenses such as an 85mm portrait or telephoto lens show less, blur the background more and make distant things look closer together. A macro lens is for very small subjects up close. Treat these as hints about the look, not exact camera settings.
Lighting
- Golden hour: low, warm sun just after sunrise or before sunset.
- Blue hour: the cool twilight after sunset.
- Backlit: the light behind the subject, leaving a bright outline.
- Overcast: soft light with faint shadows.
Frequently asked questions
What makes a good AI image prompt?
Say what the picture is of, where it is, and how it should look: the medium (photo, watercolor, 3D render), the light, how close the camera is and from what angle, the colors and the shape of the image. OpenAI’s and Google’s own guides both recommend describing these visible details specifically rather than relying on a single mood word. The builder puts those pieces in a clear order so none is forgotten.
Why is the prompt different for ChatGPT, Gemini and Midjourney?
Each one documents its own way of handling a few things. Midjourney takes the aspect ratio and the things to leave out as parameters at the end (--ar and --no). OpenAI suggests stating the aspect ratio and the exclusions in the prompt itself. Google’s guide gives a sentence template and advises describing what you want instead of writing “no …”. The builder follows each one’s documentation, linked under the prompt.
Is the picture next to the prompt made by AI?
No. It is a simple drawing that changes with your choices, so you can see what words like “low angle”, “close-up” or “backlit” mean before you use them. No image is generated on this page, and real results from any generator will look different.
Will the generator follow every part of the prompt?
Not always. A prompt describes what you want; the model decides how to draw it, and results vary between runs and models. If something is ignored, try moving it earlier, describing it in more detail or generating again.
Is what I type sent anywhere?
No. The prompt is put together in your browser. Your choices are kept in the page address so you can bookmark or share them, and nothing is sent to an image generator from here.
Can I use the prompts for commercial work?
The prompt text is yours to use. Whether you can use the images commercially depends on the terms of the image generator you use and your plan with it, so check those terms.
Did this tool do what you needed?
Related tools
Answer four questions and see which ChatGPT, Claude or Gemini plan fits, with reasons.
ChatGPT, Claude and Gemini paid plans side by side: price, limits and features.
Paste text from ChatGPT, Claude or Gemini and get clean plain text back.