A Sora prompt is a short shot description (subject, action and setting) plus camera direction: shot type, camera movement, lens, lighting, style and, for models with audio, dialogue and sound. This free generator turns your idea into a ready-to-paste prompt for Sora 2, Veo 3, Runway, Kling, Luma or Pika, with a storyboard and three variations, all in your browser.
OpenAI Sora 2 prompt
62 wordsA golden retriever puppy chases a red kite across the sand in a windy seaside cove at sunset. Cinematic film look, joyful mood. Cinematography: Camera shot: wide establishing shot, 35mm lens Camera motion: tracking shot following the subject Lighting: golden hour sunlight Mood: joyful, steady pacing Actions: - chases a red kite across the sand Background sound: waves, wind and happy barking
Set in the app: 16:9 aspect ratio, 8 seconds, sound on.
OpenAI Sora 2 has no negative prompt box, so these are left out of the prompt: naming them can add them. Describe the opposite instead, for example "clean frame" or "sharp focus". Avoiding: text or captions, watermark, distorted faces.
OpenAI Sora 2 tip: One camera move and one clear action per shot. Shorter clips follow the prompt more closely. Set length and size in the app, because the prompt cannot change them.
Generate each shot as its own clip. Every shot repeats the same subject, setting and look, which keeps characters and places consistent between generations.
Shot 1 (3s, wide establishing shot, slow dolly in): Chases a red kite across the sand. Subject: a golden retriever puppy. Setting: a windy seaside cove at sunset. Look: cinematic film look, golden hour sunlight, joyful mood. Shot 2 (3s, medium shot, tracking shot following the subject): Chases a red kite across the sand. Subject: a golden retriever puppy. Setting: a windy seaside cove at sunset. Look: cinematic film look, golden hour sunlight, joyful mood. Shot 3 (3s, close-up, static locked-off camera): Chases a red kite across the sand. Subject: a golden retriever puppy. Setting: a windy seaside cove at sunset. Look: cinematic film look, golden hour sunlight, joyful mood.
Each variation changes one thing, so you can see which choice made the difference.
A golden retriever puppy chases a red kite across the sand in a windy seaside cove at sunset. Cinematic film look, joyful mood. Cinematography: Camera shot: wide establishing shot, 35mm lens Camera motion: aerial drone flyover Lighting: golden hour sunlight Mood: joyful, steady pacing Actions: - chases a red kite across the sand Background sound: waves, wind and happy barking
A golden retriever puppy chases a red kite across the sand in a windy seaside cove at sunset. Cinematic film look, joyful mood. Cinematography: Camera shot: wide establishing shot, 35mm lens Camera motion: tracking shot following the subject Lighting: soft overcast daylight Mood: joyful, steady pacing Actions: - chases a red kite across the sand Background sound: waves, wind and happy barking
A golden retriever puppy chases a red kite across the sand in a windy seaside cove at sunset. Photoreal commercial, joyful mood. Cinematography: Camera shot: wide establishing shot, 35mm lens Camera motion: tracking shot following the subject Lighting: golden hour sunlight Mood: joyful, steady pacing Actions: - chases a red kite across the sand Background sound: waves, wind and happy barking
A glass perfume bottle on black marble turns slowly as mist drifts across it in a dark studio. Photoreal commercial, mysterious mood. Cinematography: Camera shot: close-up, macro lens Camera motion: slow orbit around the subject Lighting: moody low-key lighting Mood: mysterious, slow and calm pacing Actions: - turns slowly as mist drifts across it Background sound: soft whoosh of mist; music: minimal ambient synth pad
A white running sneaker lands in a shallow puddle and throws up a crown of water in a wet city street at night. Photoreal commercial, energetic mood. Cinematography: Camera shot: low-angle shot, 50mm lens Camera motion: static locked-off camera Lighting: neon night lighting Mood: energetic, fast and energetic pacing Actions: - lands in a shallow puddle and throws up a crown of water Background sound: splash and distant traffic; music: punchy electronic beat
A young woman in a hoodie holding her phone at arm's length talks to the camera while making coffee in a small sunny kitchen. UGC smartphone video, cozy mood. Cinematography: Camera shot: medium close-up, 24mm wide lens Camera motion: handheld camera Lighting: soft overcast daylight Mood: cozy, steady pacing Actions: - talks to the camera while making coffee Dialogue: - A young woman in a hoodie holding her phone at arm's length: "Okay, this is the one thing that fixed my mornings" Background sound: kettle boiling, room tone
A man at a desk opens a small cardboard box and lifts out wireless earbuds, grinning in a home office with a ring light. UGC smartphone video, joyful mood. Cinematography: Camera shot: medium shot, 35mm lens Camera motion: static locked-off camera Lighting: bright studio softbox lighting Mood: joyful, fast and energetic pacing Actions: - opens a small cardboard box and lifts out wireless earbuds, grinning Dialogue: - A man at a desk: "No way, they actually fit in the case!" Background sound: cardboard rustling
A red vintage car drives along a winding mountain road in misty alpine peaks at dawn. Cinematic film look, epic mood. Cinematography: Camera shot: extreme wide shot Camera motion: aerial drone flyover Lighting: blue hour twilight Mood: epic, slow and calm pacing Actions: - drives along a winding mountain road Background sound: wind and a distant engine
Raindrops on a café window run down the glass while people pass outside with umbrellas in a quiet café in the evening. 35mm film grain, vintage, melancholic mood. Cinematography: Camera shot: close-up, 85mm portrait lens with shallow depth of field Camera motion: slow dolly in Lighting: practical lamps in a dim room Mood: melancholic, slow and calm pacing Actions: - run down the glass while people pass outside with umbrellas Background sound: rain on glass, quiet café chatter; music: soft jazz piano
A bean seed in clear soil splits open, sends down a root and pushes a green shoot up in a cutaway glass planter. Documentary realism, calm mood. Cinematography: Camera shot: extreme close-up, macro lens Camera motion: static locked-off camera Lighting: bright studio softbox lighting Mood: calm, steady pacing Actions: - splits open, sends down a root and pushes a green shoot up Background sound: music: gentle curious piano
A small round friendly robot points at a floating diagram of the water cycle in a bright classroom. Stylized 3D animation, joyful mood. Cinematography: Camera shot: medium shot, 35mm lens Camera motion: slow dolly in Lighting: soft overcast daylight Mood: joyful, steady pacing Actions: - points at a floating diagram of the water cycle Dialogue: - A small round friendly robot: "Water goes up as vapor and comes back down as rain!" Background sound: light classroom room tone
A singer in a long coat sings into a vintage microphone as wind moves her hair in a city rooftop at night. Cinematic film look, dreamy mood. Cinematography: Camera shot: medium close-up, anamorphic lens with soft flares Camera motion: slow orbit around the subject Lighting: neon night lighting Mood: dreamy, slow and calm pacing Actions: - sings into a vintage microphone as wind moves her hair Background sound: music: slow synth-pop ballad
Two paper cut-out dancers spin and leap across a sheet of music in a wooden desk. Claymation stop-motion, joyful mood. Cinematography: Camera shot: high-angle shot Camera motion: crane up Lighting: warm candlelight Mood: joyful, fast and energetic pacing Actions: - spin and leap across a sheet of music Background sound: music: upbeat swing with clarinet
| Model | Prompt structure | Sound | Negative prompt | Tip |
|---|---|---|---|---|
| OpenAI Sora 2 | A short prose scene, then labeled Cinematography, Actions, Dialogue and Background sound blocks. | Yes, generated with the video | None; describe positives | One camera move and one clear action per shot. Shorter clips follow the prompt more closely. Set length and size in the app, because the prompt cannot change them. |
| Google Veo 3 | One descriptive paragraph: shot and camera, subject, action, setting, lighting and style, then dialogue in quotes and sound cues. | Yes, generated with the video | Separate field | Put spoken lines in quotes after the speaker. In the negative prompt, list what to leave out as plain nouns ('text, watermark'), not 'no text'. |
| Runway Gen-4 | Camera move first, then the scene and the motion, in plain direct sentences. | Silent in most versions | None; describe positives | Describe motion, not the look of a still image (the reference image covers that). Avoid negative phrasing: it can add the thing you named. |
| Kling | Subject and details, movement, scene, then camera language, lighting and atmosphere. | Silent in most versions | Separate field | Name the subject's motion with a strong verb and keep one main subject. Use the separate negative prompt box for artifacts. |
| Luma Dream Machine (Ray) | Natural, conversational sentences that say what happens and how the camera moves. | Silent in most versions | None; describe positives | Write it like you are describing the shot to a camera operator. Add a camera move in plain words, such as 'the camera slowly pushes in'. |
| Pika | One short line: subject, action, setting, camera and style. | Silent in most versions | Separate field | Keep it short and visual. Put artifacts you want to avoid in the negative prompt field. |
Models change often. Check each app for its current clip lengths, sizes and audio support.
Models lose track when a single shot has several people doing several things. Split busy scenes into a storyboard.
'Slow dolly in' beats 'dolly in, then orbit, then crane up'. Combine moves across shots, not inside one.
Lighting changes the result more than adjectives like 'beautiful'. Say 'golden hour sunlight' or 'neon night lighting'.
Words like 'no people' can add people. Describe what you want ('an empty street'), and use a negative prompt box where there is one.
Plan about one distinct action per three seconds. An 8-second clip fits two actions comfortably.
Change only the camera, the light or the style between runs, so you know what made the clip better.
AI video apps block violent, sexual and hateful content, and most refuse prompts that show real public figures or a real person without consent. Sora only puts a real person in a video through that person's own opt-in cameo. Do not upload or describe someone's likeness without permission, and label AI video as AI-made when you share it.
A good Sora prompt names one subject, one clear action and a setting, then gives the shot type, one camera move, the lighting and the look. Keep each shot to one or two actions, and put spoken lines and sounds in their own lines. Set clip length and size in the app, because the prompt cannot change them.
Yes. Sora 2 and Veo 3 can generate dialogue, sound effects and ambient sound with the video, so the generator adds a dialogue line and sound cues for them. Runway, Kling, Luma and Pika versions without native audio make silent clips; add music and sound in an editor.
Usually 40 to 120 words. Sora and Veo handle longer, structured prompts well; Pika and Luma do best with a short, visual line. Longer is not better if it adds more actions than the clip has time for: plan about one distinct action per three seconds.
Describe the character, setting and look the same way, word for word, in every shot, and change only the framing and action. The storyboard output does this for you. Using the same reference image or a saved character, where the model offers one, helps even more.
No. It is a free, template-based builder that runs in your browser with no signup and no AI API calls. Nothing you type is sent to a server. You paste the prompt into Sora, Veo, Runway, Kling, Luma or Pika yourself.
Sora is a trademark of OpenAI; Veo of Google; Runway, Kling, Luma Dream Machine and Pika of their owners. AI Wins is not affiliated with or endorsed by any of them.
Make a still first frame or reference image in Midjourney, DALL-E or Stable Diffusion.
Open toolScore your clip with an AI song: style tags, lyrics metatags and structure.
Open toolBorrow a visual style for your video with copyable prompt keywords.
Open toolDraft a script or shot ideas with a structured ChatGPT or Claude prompt.
Open toolNew AI video models launch every few weeks. Follow the launches and the wins in the free daily AI Wins digest.
Get the free digestOnly positive AI stories, delivered daily or weekly. Free, no spam, unsubscribe anytime.