AI Video Prompt Generator: Shot Builder
Describe one shot, choose a camera move and copy a finished paragraph. Switch to image-to-video when your starting image already supplies the subject and look.
FreeNo signupRuns in your browser
Your starting image fixes the subject, setting, style and lighting. Describe only motion, camera and sound here.
One visible action in the present tense, such as shapes a bowl or turns toward the window.
Include the location phrase, such as in a studio or beside a river.
A requested length. Check the settings in your video tool.
Writing the campaign around your clip? Browse free copy-ready text prompts for marketing and SEO, sales, coding and careers.
Browse free copy-ready prompts →Camera movement glossary
Choose the move that serves the action. A camera instruction should tell you how the view changes over the shot, not just that the result should look cinematic. Use one move at a time so you can judge whether the model followed it.
| Move | What it looks like | When to use it |
|---|---|---|
| Static | The camera stays in place. | Let a gesture or a moving object carry the shot. |
| Slow push in | The camera moves closer to the subject. | Draw attention to an expression or small detail. |
| Pull back | The camera moves away from the subject. | Reveal more of the surroundings. |
| Pan left or right | The camera turns horizontally from one position. | Follow movement across a scene. |
| Tilt up or down | The camera turns vertically from one position. | Reveal the height of a subject or a detail below it. |
| Orbit | The camera travels around the subject. | Show a product or sculpture from changing angles. |
| Tracking shot | The camera travels with a moving subject. | Stay with a walker, cyclist or vehicle. |
| Handheld | The framing has small operator-like movements. | Request an observational, less fixed viewpoint. |
| Crane up | The camera rises above its starting position. | Reveal the space around the subject. |
| Drone flyover | An aerial view travels over the setting. | Establish a landscape or a route. |
Text-to-video versus image-to-video prompts
For text-to-video, the prompt needs to establish what is visible. Describe the subject, its action and the setting before adding the look. Models read the start of the prompt most strongly, so put the shot you actually want before decorative details.
For image-to-video, your source image already fixes appearance. Repeating clothing, lighting and style wastes words and can cause drift. Switch modes to hide those fields and focus on what changes: motion, camera and sound. Your hidden values are kept if you switch back. Use the image prompt generator to plan a starting image, then follow the guide to effective image-to-video prompts when describing its movement.
How to use the shot builder
- Choose your starting point and replace the example with one subject and one visible action.
- Set the shot size and one camera movement. Leave a choice unspecified when it does not matter to the shot.
- Add sound if you want it, then give any dialogue a speaker. Models that generate audio use these cues; silent models ignore them.
- Read the paragraph aloud, copy it and choose suitable settings in your video tool. The requested duration is part of the brief, not a model limit.
The builder assembles text locally. It does not generate footage or check your chosen model's capabilities. If you want to keep the same brief as editable fields, use the JSON prompt generator.
Common failures and the fix
Too many actions: a character who enters, sits, opens a box and leaves needs several shots. Pick the moment that matters and make the other moments separate prompts. The status reminder flags sequencing words and long comma lists, but you should still review the action yourself.
Vague motion: replace “moves dramatically” with “turns toward the window.” A visible verb gives you something concrete to assess. Keep dialogue short enough to say within a clip of a few seconds; long lines can get cut off or rushed.
Conflicting camera moves: remove a second move from your action or setting if you already chose one in the camera field. Change one instruction after a poor result so you can tell which edit helped.
Frequently asked questions
How do I write a good AI video prompt?
Describe one subject, one action and one camera move. Put the main shot first, then add lighting, speed and sound cues. This builder assembles those choices into a paragraph you can copy.
How is an image-to-video prompt different?
An image-to-video prompt should focus on motion, camera and sound. The source image already controls appearance, so repeating the subject, setting, lighting and style wastes words and can cause drift. Those fields are hidden in image-to-video mode.
Can I add dialogue and sound to a video prompt?
Yes, use the ambient sound and dialogue fields. Models that generate audio can use these cues, while silent models ignore them. Give the speaker a name or description so the dialogue has a clear source.
How long should dialogue be in a short AI video?
Keep dialogue to a short line that fits the requested clip. Long lines can be rushed or cut off. Read the line aloud before copying the prompt and shorten it if it takes the whole clip to say.
Does this free video prompt generator create a video?
No, it builds the text of a shot description in your browser. Copy that text into your video tool and choose the available generation settings there. The duration in the prompt is a request, not a guarantee of supported clip length.
