Free prompt tool

Video Prompt Generator for Sora, Veo, Runway, and Kling

Convert a rough creative idea into a structured video prompt with subject continuity, camera movement, timing beats, reference image handling, and a negative prompt.

Build a video prompt

Fill the structure once, then tune the model target and copy the output into Sora, Veo, Runway, Kling, or another video generator.

Target modelAspect ratio

SubjectSettingKey outcome

Camera moveDuration

Visual styleReference image handlingAudio or dialogue notesNegative prompt

Generated prompt

Create a 10 seconds 16:9 video for Sora.
Scene: a compact studio with soft daylight, a laptop, printed sketches, and a clean product display.
Subject: a designer reviewing AI-generated product mockups on a desk.
Goal: the rough concept becomes a polished launch visual on the screen.
Visual style: realistic product demo, natural color, crisp detail, no glossy commercial exaggeration.
Camera: slow push-in from waist height; keep the motion smooth and physically plausible.
Reference handling: Keep the subject identity, layout, and brand colors consistent with the reference image.
Audio: Subtle room tone only. No voiceover unless the target model supports audio.
Continuity: preserve the same subject, lighting direction, object positions, and screen content across the full clip.
Negative prompt: warped hands, unreadable text, melting objects, sudden cuts, flicker, camera shake, extra fingers, distorted logos.

Shot list

0-2s: Establish a designer reviewing AI-generated product mockups on a desk in the setting with a clean readable composition.
2-5s: Begin slow push-in from waist height while the main action starts.
5-8s: Show the key transformation or proof point: the rough concept becomes a polished launch visual on the screen.
8-10s: Hold on the finished state for a usable final frame.

Sora tuning notes

TipPut the continuity rule near the top because Sora-style prompts benefit from stable subject identity and spatial constraints.

TipDescribe camera motion as a physical move, not just a vibe.

TipUse a short ending state so the clip has a clear final frame.

Model-specific prompt shape

The same idea is reorganized for Sora, Veo, Runway, Kling, or a generic model so the strongest instruction lands first.

Shot list instead of vague style

The output includes time beats so motion, reveal, and final frame are easier to inspect after generation.

Reference image control

Image-to-video prompts get explicit rules for what must remain stable from the source image.

How to use the generated prompt

Start with one clean generation. If the clip misses the composition, fix subject, setting, and camera before adding style. If the motion is wrong, edit the shot list and remove abstract adjectives. Keep brand names, copyrighted characters, and private people out unless you have the right to use them.

Estimate image API cost Compare model status