What Makes a Good Gemini AI Video Prompt?
If you have typed Gemini AI video prompts into search, you are probably looking for complete scripts that turn a text description into a moving scene. This guide gives you a set of specific, copy ready prompts you can use in Gemini's video generation tool or in the dradra mobile app. Each prompt includes subject, action, camera movement, lighting, and style details. You can copy them exactly or change a few words to fit your own footage.
A strong video prompt gives the model a clear scene to imagine. Instead of writing 'a dog running', write 'A golden retriever runs across a shallow creek at golden hour, water splashing, camera tracking from a low side angle, warm backlight, shallow depth of field, 5 seconds, cinematic'. The extra details reduce randomness and help you get the framing, mood, and motion you want.
36 Gemini AI Video Prompts
These prompts are organized by common video styles. Copy one, paste it into your AI video tool, and edit the subject or setting to match your project.
Cinematic and Moody
- A slow push-in through a rain-streaked window, a person in a dark coat looking out at neon city lights, blue and orange contrast, gentle rain droplets on glass, shallow depth of field, 4 seconds.
- An empty roadside diner at night, one flickering neon sign, steam rising from a coffee cup, camera slowly dollies left, muted colors, film grain, 5 seconds.
- A lone figure walks down a foggy bridge at dawn, long shadows, soft gray light, camera follows from behind, atmospheric haze, 6 seconds.
- Close-up of an eye blinking slowly, reflection of a candle flame moving across the iris, dark background, macro lens, cinematic 24fps, 3 seconds.
- An old piano in a dusty room, a shaft of sunlight illuminates floating dust particles, camera tilts down from window to keys, warm nostalgic tone, 5 seconds.
- A car drives through a mountain tunnel at night, tail lights streak, overhead camera, moody blue and red lights, motion blur, 4 seconds.
Character and Portrait
- A ballet dancer spinning in an empty white studio, soft window light, camera orbits around her, fabric flowing, slow motion, 6 seconds.
- A fisherman casting a net on a wooden boat at sunrise, golden rim light, camera moves from side profile to front, sparkle on water, 5 seconds.
- A barista pouring latte art in a cozy cafe, close-up from above, steam curling, warm color grade, shallow depth of field, 4 seconds.
- An astronaut standing on a rocky alien shore looking at two moons, slow 360 pan, teal and orange color palette, epic scale, 6 seconds.
- A little girl releasing a paper boat into a puddle, reflection of clouds, camera low to the ground, gentle splashes, whimsical mood, 5 seconds.
- A musician's fingers moving across guitar strings in a dim bar, extreme close-up, bokeh lights in background, slow pull back, 4 seconds.
Nature and Wildlife
- Aerial shot flying over a winding river through autumn forest, leaves falling, camera tilts down, crisp morning light, 5 seconds.
- A hummingbird hovering near purple flowers, wings frozen in high-speed detail, macro lens, soft green background, 3 seconds.
- Ocean waves crashing against black volcanic rocks at sunset, slow motion spray, ultra wide angle, orange and deep blue, 6 seconds.
- Timelapse of a cherry blossom tree opening its flowers over spring, soft pastel pink, gentle breeze, 8 seconds.
- A red fox walking through light snow in a pine forest, camera tracks beside it, breath visible, cold blue shadows, 6 seconds.
- A waterfall in a tropical rainforest, long exposure silky water, camera slowly ascends, lush green foliage, 5 seconds.
Urban and Street
- A rainy Tokyo alley at night, neon signs reflecting on wet pavement, people with clear umbrellas walk by, camera glides forward, 5 seconds.
- A skateboarder grinding a rail in a concrete skatepark, low angle tracking shot, golden hour flare, dust particles, 4 seconds.
- Street food vendor tossing noodles in a flaming wok, high-speed close-up, sparks and steam, camera shake for energy, 4 seconds.
- A crowded subway platform, train arrives with motion blur, people step aside, fluorescent light, shallow depth of field, 5 seconds.
- A mural artist painting a large wall, cherry picker moving up, time lapse style, colorful spray paint mist, 6 seconds.
- A cyclist weaving through early morning city traffic, mounted camera perspective, motion blur, cool blue morning light, 4 seconds.
Surreal and Fantasy
- A giant whale swimming through clouds above a coastal town, soft pastel sky, camera follows its arc, dreamlike, 6 seconds.
- Floating islands with waterfalls cascading into an endless sky, camera flies between them, golden sunset, fantasy epic, 6 seconds.
- A clock tower melting into a puddle of liquid gold, close-up of dripping numbers, reflective surface, surreal, 4 seconds.
- A girl's hair flowing upward becoming a galaxy of stars, camera orbits from below, deep blues and purples, 5 seconds.
- A doorway standing alone in a desert, light pouring from inside, camera slowly pushes through, dust blowing, mysterious, 5 seconds.
- Mythical deer with glowing antlers walking through a dark forest, fireflies floating, camera follows at low angle, 6 seconds.
Product and Commercial
- A perfume bottle on a wet black surface, water ripples around it, dramatic backlight, camera orbits, tiny droplets suspended, 4 seconds.
- A smartphone rotating in midair, glowing screen showing a colorful app, dark studio background, product lighting, 5 seconds.
- A sports shoe hitting a black surface sending particles of dust and water, high-speed side view, sharp detail, 3 seconds.
- A lipstick tube opening in slow motion on a marble table, soft pink light, camera sweeps from above, elegant, 4 seconds.
- A coffee mug on a wooden table with steam forming shapes, warm morning window light, camera slowly pushes in, 5 seconds.
- A luxury watch with gears turning, macro shot, golden light, dust motes, camera rotates around the face, 5 seconds.
How to Adapt These Prompts for Your Own Project
These prompts are starting points, not fixed scripts. Replace the subject, change the environment, or swap the style keyword to match your needs. For example, turn the rainy Tokyo alley into a sunny Lisbon street by changing location and lighting. If you need vertical video for Shorts or Reels, add 'vertical 9:16' and reduce the described camera movement. If you want an anime look, add '2D anime style' or 'Studio Ghibli style' after the action. If your video tool supports a negative prompt field, add terms like 'blurry, distorted, text, watermark' to reduce unwanted artifacts.
When a result looks stiff, add explicit camera movement such as 'slow push-in', 'orbit left', or 'tracking shot'. When lighting feels flat, specify a time of day and a light source, such as 'golden hour backlight' or 'cool moonlight from the left'. When the scene looks cluttered, simplify to one subject and one background. These small changes often turn a rough draft into a usable clip.
Common Weak Results and How to Fix Them
- Stiff motion: Add a camera move and describe the subject's action in more detail.
- Flat lighting: Name the time of day, light direction, and color temperature.
- Distorted faces or hands: Keep the subject at medium distance, avoid extreme close-ups, and mention 'sharp facial details'.
- Cluttered background: Use a simpler setting and say 'uncluttered background' or 'shallow depth of field'.
- Wrong aspect ratio: Add 'vertical 9:16' or 'horizontal 16:9' at the end of the prompt.
Using These Prompts in dradra
If you want a mobile first workflow, the dradra app on Android supports text-to-video generation with over one million style filters. Copy any prompt above, open dradra, select the text-to-video tool, paste the prompt, and choose a style template such as cinematic, anime, or realistic. The same prompt structure works across many text-to-video models, so you can also run these on your preferred AI video tool.
Start with one prompt that matches your current project. Tweak a single element, such as the subject or camera angle, and generate again. Small iterations usually produce a clip that feels intentional rather than random.