A good AI image prompt names five things in one or two plain sentences: the subject, the setting, the light, the framing and the style. Add what you do not want, such as "no text", when it matters. Below are 20 copy-ready prompts, grouped by use, each shown with the image FrameSurfer's text to image model made from it.
Copy a prompt, swap in your subject and keep the structure. Fixes for common failures, shape advice and animation tips follow the examples.
The AI image prompt formula
Describe one visible moment, not a story, using these parts in roughly this order.
| Part | What to write | Example |
|---|---|---|
| Subject | The one main thing, with a detail or two | A handmade ceramic mug |
| Setting | Where it is and what is near it | On a linen cloth beside a window |
| Light | The light source and the time of day | Soft morning light |
| Framing | Camera distance, angle, or where the subject sits | Close-up, eye level, subject on the right |
| Style | One medium: a type of photo, illustration or craft | Still life photograph |
| Leave out | Things you do not want in the picture | No text, no logos, no people |
Prompt 1 below uses four of them. You rarely need all six. If you add only three, pick the subject, the light and the style.
Use specific nouns, not praise words. "Linen cloth" tells the model something. "Beautiful" does not.
20 AI image prompts with the images they made
Each example shows the aspect ratio, the exact prompt and its image. Paste any prompt into FrameSurfer's Text to Image tool and pick the same shape. Each run is a fresh generation, so your image will differ in the details.
Product shots
1. Ceramic mug by a window (vertical)
A handmade ceramic mug on a linen cloth beside a window, soft morning light, still life photograph, shallow depth of field.
FrameSurfer Text to Image, from the prompt above.
Why it works: two materials, one light source, and "shallow depth of field" (a blurred background) so the mug stands out.
2. Serum bottle on wet slate (1:1)
An amber glass dropper bottle standing on wet slate, water droplets on the glass, soft studio light from the left, dark green background, close-up product photograph, blank label, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: "blank label, no text" asks for no fake lettering. One light direction and one background color read as a studio shot.
3. Leather backpack in a doorway (9:16)
A tan leather backpack hanging on a wooden peg beside a sunny doorway, a folded wool jacket on the bench below, warm afternoon light, lifestyle product photograph at eye level, no logos.
FrameSurfer Text to Image, from the prompt above.
Why it works: the jacket and doorway show the product in use without showing a person.
Food and drink
4. Iced coffee on a cafe counter (1:1)
Iced coffee in a tall glass on a marble cafe counter, milk swirling into the coffee, condensation on the glass, bright morning window light, close-up food photograph, shallow depth of field.
FrameSurfer Text to Image, from the prompt above.
Why it works: swirling milk catches a moment, and condensation tells you the drink is cold.
5. Breakfast table from above (16:9)
Top-down view of a rustic breakfast table with a sourdough loaf, a bowl of berries, a small jar of honey and two cups of tea on worn wood, soft overcast daylight, food flat lay photograph.
FrameSurfer Text to Image, from the prompt above.
Why it works: "top-down" and "flat lay" (shot from directly above) fix the angle. Naming each item controls what sits on the table.
Portraits of made-up people
These are invented people. Describe age, job, clothing and setting, never a real or famous name.
6. Baker at dawn (9:16)
Chest-up portrait of a baker in her fifties with flour on her apron, standing in a small bakery at dawn, warm light from the ovens behind her, calm smile, 85mm portrait photograph.
FrameSurfer Text to Image, from the prompt above.
Why it works: one detail (flour) makes a believable person. "Chest-up" keeps hands, a common weak spot, out of frame. "85mm" is a classic portrait lens.
7. Fisherman at sunrise (1:1)
Close-up portrait of an older fisherman with a grey beard and a knitted cap on a harbor at sunrise, sea spray in the air, soft side light, weathered skin in sharp detail, documentary photograph.
FrameSurfer Text to Image, from the prompt above.
Why it works: side light shows texture, and "documentary photograph" asks for a natural, unretouched look.
Need the same face again? Text to Image invents a new person each run. Character Portrait makes one front-facing portrait you can save and reuse.
Interiors and real estate
8. Scandinavian living room (16:9)
A bright Scandinavian living room with a light oak floor, a linen sofa, a large window and green plants, late morning sun, wide-angle interior photograph taken from the doorway, tidy and uncluttered.
FrameSurfer Text to Image, from the prompt above.
Why it works: naming the lens (wide-angle) and where the camera stands (the doorway) gives a real interior-photo view.
9. Green kitchen (9:16)
A small modern kitchen with dark green cabinets, a brass tap and a white stone counter, a bowl of lemons by the sink, soft daylight from a side window, vertical interior photograph, no people.
FrameSurfer Text to Image, from the prompt above.
Why it works: one color story (green, brass, white) and one prop look styled, not busy.
10. Timber house at dusk (16:9)
The front of a two-story timber house at dusk, warm light glowing in every window, a neat lawn and a stone path, deep blue sky, real estate exterior photograph at eye level, wide shot.
FrameSurfer Text to Image, from the prompt above.
Why it works: a house with every window lit against a deep blue dusk sky is the classic real estate twilight shot.
Use generated rooms for mood boards, concepts and backgrounds. Never present one as a real property.
Travel, cities and scenery
11. Neon diner in the rain (vertical)
A neon-lit diner window on a rainy night, reflections on wet asphalt, photorealistic, moody, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: weather and reflections carry the mood. "No text" matters with neon, and this image came back with no lettering.
12. Fishing village from above (16:9)
A small whitewashed fishing village on a rocky coast, blue boats in the harbor, turquoise water, late afternoon sun, aerial view from a drone, travel photograph.
FrameSurfer Text to Image, from the prompt above.
Why it works: "aerial view from a drone" sets the viewpoint, and three named colors keep the palette clean.
13. Alpine lake at sunrise (9:16)
A calm alpine lake at sunrise, mist over the water, pine forest and snowy peaks reflected in the surface, one wooden rowboat at the shore, cool blue and pink light, vertical travel photograph.
FrameSurfer Text to Image, from the prompt above.
Why it works: the single rowboat gives a quiet scene a focal point.
Illustration styles
14. Watercolor fox (1:1)
A red fox curled up asleep in fresh snow, loose watercolor illustration on textured paper, soft washes of orange and blue, plenty of white paper around it, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: it names the medium and the paper, and the white space keeps watercolor's soft edges.
15. Flat illustration of a cyclist (16:9)
Flat vector illustration of a person riding a bicycle through a city park in autumn, simple shapes, a limited palette of mustard, teal and cream, no outlines, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: three named colors make a tight palette, and "no outlines" keeps it clean for headers or slides.
16. Paper-cut lighthouse (1:1)
A layered paper-cut scene of a lighthouse on a cliff at night, a full moon and rolling waves, each paper layer casting a soft shadow, deep blue and white paper, photographed straight on, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: describing layers that cast shadows makes it read as real paper, not a flat drawing.
Social backgrounds and thumbnails
17. Background with room for text (9:16)
A soft peach background with a few dried flowers and a folded linen napkin in the bottom corner, empty space across the upper two thirds, gentle natural shadows, minimal flat lay photograph, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: it says where the empty space goes, so your caption has room.
18. Thumbnail-style close-up (16:9)
A golden retriever wearing small round glasses, sitting at a desk piled with books and looking at the camera in surprise, bold orange background, bright studio light, dog on the right side of the frame, empty space on the left, no text.
FrameSurfer Text to Image, from the prompt above.
Why it works: one bold color, a clear expression and an empty left side for a title read well at thumbnail size.
Stills meant to be animated later
If the still will become a clip, include something that can move: dust, water, light or fabric.
19. Trail runner on a ridge (vertical)
A trail runner cresting a ridge at golden hour, dust in the air, wide cinematic still, warm backlight.
FrameSurfer Text to Image, from the prompt above.
Why it works: dust and a clear direction of travel give a video model something to move. Note the shape: the prompt says "wide", yet the image came back tall. The aspect ratio setting, not the words, sets the frame.
20. Lanterns on a river (9:16)
Hundreds of glowing paper lanterns floating on a quiet river at dusk, a wooden footbridge in the background, warm orange light against a deep blue sky, cinematic still, no people.
FrameSurfer Text to Image, from the prompt above.
Why it works: lanterns and water are made for motion, and with no people there are no faces or hands to go wrong.
Common failures and how to fix them
Most weak results come from a vague or overloaded prompt. Change one thing at a time so you can tell what fixed it.
| What goes wrong | Likely cause | Try this |
|---|---|---|
| Hands with extra or bent fingers | Hands in a busy pose, or holding things | Frame chest-up, or give the hands a simple pose such as "hands in pockets" |
| Misspelled words or fake logos | The prompt asks for words, or the scene invites them (signs, labels, book covers) | Add "no text" and "blank label", then add real words in an editor |
| A cluttered, confusing scene | Several subjects competing | One main subject and two or three props at most |
| Flat, dull light | Vague words like "good lighting" | Name the source and time: "window light from the left", "golden hour backlight" |
| A mix of looks | Several styles in one prompt | Pick one medium: photograph, watercolor or flat illustration |
| The subject is too small | No framing words | Add "close-up", "chest-up" or "fills the frame" |
| No room for a headline | Empty space was never requested | Name the empty side, as in prompt 18 |
Save a version you like first. A new run can change other details too.
Pick the shape before you write
Text to Image offers three aspect ratios. Aspect ratio is the shape (width compared with height), not the resolution.
| Aspect ratio | Shape | Good for |
|---|---|---|
| 9:16 | Tall | Stories, Reels, Shorts and TikTok, or a vertical video frame |
| 1:1 | Square | Feed posts, product tiles, profile images |
| 16:9 | Wide | YouTube thumbnails, blog headers, widescreen video scenes |
For thumbnails, YouTube's help page suggests 16:9 for videos and 9:16 for Shorts, as of September 2026 (YouTube Help).
When you change the shape, rewrite the framing too. A wide shot with the subject on the right becomes a tall one with the subject low and sky above.
Turn a still into a clip: image to video prompts
Image to Video uses your still as the first frame and animates it from a line about motion. The rule: describe the change, not the picture. Describing only motion helps the subject stay as it is.
Here is a motion prompt for image 20:
The lanterns drift slowly downstream, their reflections ripple on the water, and the camera pushes gently toward the bridge. Nothing else moves.
It follows the tool's own examples: one camera move, one or two changes, and "nothing else moves" to hold the rest still.
You get one MP4 in 9:16, 1:1 or 16:9, with generated sound on or off. Maximum length depends on the video model, from 8 seconds on Veo 3.1 Fast to 30 seconds on Alibaba Wan 3. See how to animate a picture for a walkthrough and the AI video prompts guide for more motion wording.
How to use Text to Image in FrameSurfer
- Write the prompt. Open Text to Image and describe what should be visible in the Image prompt box.
- Choose the aspect ratio. Pick 9:16, 1:1 or 16:9 for where the image will appear.
- Select Generate. Writing the prompt needs no account. Generating needs a FrameSurfer plan, and after sign-up the tool opens with your prompt in it.
- Review it. Check it at the size people will see it, then zoom in on hands and small details. Each run makes one image, so change the prompt for another take.
- Use it. Download it, or take it into Image to Video as the first frame.
Start from a photo when the real thing matters
Text to Image invents everything in the frame. That suits concepts and backgrounds, not pictures of your actual product. Product Photo starts from your product photo and places it in a new scene, keeping it recognizable. The AI product photography guide covers that workflow.
Copy any prompt above into FrameSurfer's Text to Image, pick the shape and generate.
Frequently Asked Questions
What makes a good AI image prompt?
It names the subject, setting, light, framing and one style in one or two sentences. Specific nouns like "wet slate" help more than praise words like "stunning".
How long should a text to image prompt be?
One or two sentences is enough for most images. Every prompt in this guide is under 50 words.
Can AI image generators put readable text in an image?
Letters can come out misspelled or garbled. Ask for "no text", then add the words in an editor where you can check them.
How many images does FrameSurfer's Text to Image make per run?
One image per run. For a different take, change the prompt and run it again.
Can I use the images commercially?
Yes. Under FrameSurfer's terms you own what you create and can use it personally or commercially.
What is the difference between an image prompt and an image to video prompt?
An image prompt describes the whole picture. An image to video prompt describes only the motion, because your still is already the first frame.
