Video to Prompt

1Video2 s to 2 min

Example

What to focus onOptional

Prompt style

Examples

Original

Prompt for shot 2 (3.3–7.3s)

Eye-level studio shot of premium beige wireless over-ear headphones resting on a dark marble pedestal, dramatic golden rim light, sleek commercial aesthetic. Headphones rotate slowly atop a polished dark marble plinth under warm rim light. Camera: slow orbit around the subject. Style: Sleek commercial aesthetic with dark backgrounds, warm rim lighting, and elegant tones. Slow, unhurried pace.

Made from the prompt

Headphone ad · 5 shots

What is Video to Prompt?

It watches a video and writes the prompts you would need to make something like it with an AI video model. You get the structure and rhythm of the whole clip, its overall look, and for every shot the framing, camera move, subject, action, setting and light.

Each shot becomes a ready-to-use prompt, and the whole clip becomes one multi-shot prompt. Switch the style to match the model you use, edit the words, then copy them or make the video right here.

What you get

Shot list. Every cut found automatically, with its time range and a thumbnail you can click to jump to.

Camera language. Shot size (wide, medium, close-up…) and camera move (push-in, orbit, tracking, handheld…) for each shot.

Ready prompts. One prompt per shot plus a multi-shot prompt for the whole clip, written in English, which video models follow best.

Structure and look. A one-line read of the hook, build and payoff, and the overall visual style.

Prompt styles for each model

Universal: scene, action, camera, style and pacing in one paragraph. Works with most models.

Kling: subject and movement first, then scene, camera language, light and atmosphere, the order Kling recommends.

Seedance: one line per shot, marked [Shot 1], [Shot 2], for multi-shot clips.

MiniMax: camera commands in brackets such as [Push in] or [Pan right] at the start of each shot.

Veo: a cinematic description with shot size, subject, action, setting, camera, lighting and style.

Sora: a scene description, then cinematography and actions in separate blocks, with timings for each shot.

What people use it for

Remake a trend. Get the camera moves and pacing of a viral clip, then make your own version with your subject.

Learn prompting. See how a great shot is described, and borrow the words for framing, light and movement.

Ads and product videos. Break down an ad you like into shots and recreate the structure for your product.

Storyboards. Turn reference footage into a shot list before you start a film in the AI film maker.

Tips for better prompts

Shorter clips give more detail per shot. Trim long videos to the part you care about.

Use the focus box to point at what matters, like the camera work or the product.

Replace the subject with your own (a person, product or pet) and keep the camera and light words.

Prompts describe people by role and clothing only, and leave out names, brands and on-screen text.

Questions

Is Video to Prompt free?

The examples are free to view. Reading your own video costs a few credits, shown on the button first, and switching styles or copying prompts never costs anything.

How long can my video be?

From 2 seconds up to 2 minutes. Up to 12 shots are described; very fast edits are grouped together.

Can I paste a link to a video?

Not yet. Download the clip you have the rights to use, then upload the file.

Which models do the prompts work with?

Any text-to-video model. The Kling, Seedance and MiniMax styles follow each model's own prompt conventions, and Veo and Sora styles are there for copying. Make this video uses MiniMax H3 by default, and you can switch models on the next page.

Will the new video look exactly like the original?

No. The prompts capture the framing, movement, light and style, not anyone's identity or exact frames, so you get a new video in the same spirit.

Does it describe people or brands?

People are described only by role, hair and clothing, never by name, age or body. Brands, logos and on-screen text are left out of the prompts.

What's the difference from Image to Prompt?

Image to Prompt reads one still. Video to Prompt finds every cut in a clip and adds what only video has: camera moves, action and pacing.