MiniMaxVion

AI Video Prompts

Real text to video and image to video prompts, each with the clip it produced. Copy one, change the subject, and generate your own shot.

Short AI clips work best when the prompt describes one moment: one subject, one action, one camera move. The text to video prompts below set the scene like a photo and then add a small amount of motion. The image to video prompts do the opposite: they lock what is already in the frame and describe only what moves.

  1. Copy a prompt

    Pick a text to video or image to video example.

  2. Change the subject

    Keep the camera and motion words, which do most of the work.

  3. Generate and compare

    Run two or three versions and keep the steadiest one.

Rainy neon street tracking shot

Cinematic tracking shot along a narrow city street at night in the rain. Neon signs glow in pink and blue without readable letters, reflections shimmer on the wet pavement, a person with a clear umbrella walks away from the camera, light steam rising from a vent. Slow steady camera, moody atmosphere. No text.

Drone flight over sea cliffs

Aerial drone shot gliding slowly over dramatic sea cliffs at sunrise. Waves crash against the dark rocks below, green grass along the cliff tops, soft golden light and low mist over the ocean. Smooth forward camera movement, cinematic and calm. No text.

Slow-motion latte art pour

Slow-motion close-up of milk being poured into a cup of black espresso, forming a swirling latte art rosetta. Warm café light, a dark ceramic cup on a wooden counter, blurred coffee bar in the background, gentle steam. Static camera, smooth motion. No text.

Ink blooming in water

Macro shot of drops of deep blue and gold ink falling into clear water, blooming into slow swirling clouds that twist and fold into each other. Pure white background, soft even light, extremely smooth slow motion. Abstract and elegant. No text.

Glass skyline at dusk

A glass skyline across mirror-still water at dusk. A tall needle spire stands in the center, the sky fades from pink at the horizon to deep blue overhead, a small moon sits in the upper right, and the towers reflect in the water. Vertical frame, gentle ripples only, no text.

Perfume shot animated

Keep the perfume bottle fixed and sharp. The water around the base ripples outward, the splash settles in slow motion, droplets fall back into the water, and a thin streak of light slowly sweeps across the glass. Keep the framing. No camera move.

Hiker in morning mist

The hiker stays on the mossy foreground rock, back to camera, dark jacket and backpack, hands in pockets. Sunbeams from the upper left shift slowly through the clouds, and mist drifts in the forested valley. Keep the framing. No camera move.

Floating islands come alive

Keep the floating islands, bridges and temple exactly in place. The clouds drift slowly below, the waterfalls flow and mist rises from where they fall, birds glide across the distance, and the sunrise light gently brightens. Keep the framing. No camera move.

Golden-hour video regrade

Keep the two hands and the open film clapperboard with the blank slate. Warm the desert light to a deeper golden hour and hold the same framing, with the dunes soft behind the board. Do not add readable text.

Text to video: one moment, one move

Each text to video prompt here describes a single continuous moment and names the camera explicitly: "slow steady camera", "smooth forward camera movement", "static camera". Leaving the camera out lets the model guess, which often gives jittery or mixed moves.

Image to video: lock first, then move

The image already shows the subject, so these prompts start with what must stay — "keep the bottle fixed and sharp", "keep the islands exactly in place" — and only then list one gentle motion per element. "Keep the framing. No camera move." at the end stops unwanted zooms that warp the subject.

Words that help

  • Camera — static, slow push-in, tracking, aerial glide
  • Speed — slow motion, gentle, smooth
  • Stability — keep the framing, no camera move, fixed and sharp
  • Exclusions — no text, no readable letters on signs

What makes a good AI video prompt?

One subject, one action and one named camera move, plus the light and mood. Short clips go wrong when a prompt tries to tell a whole story.

How is an image to video prompt different?

It describes only what moves. Start by saying what must stay fixed, then add one motion per element and keep the camera still.

Why do the prompts say "no readable letters"?

Video models struggle with text, so signs and labels often come out garbled. Asking for no readable letters avoids it.

Which model made these clips?

They were made with the Seedance video models in MiniMaxVion. Seedance 2.0 makes 4 to 15 second clips and Seedance 2.5 up to 30 seconds. Open a case to see its settings.

Can I use these prompts commercially?

You can reuse the prompt text freely. Check the terms for the videos you generate and make sure your subject and brand use are cleared.

Pricing · Text to Image · Image to Video · 中文