One line in · A shot-by-shot brief out · Render it here or take it with you
Sora 2 Prompt Generator
Describe the clip you want in one sentence and get back a full brief: the shot list, what the camera is doing in each beat, the lens and grade, and how the seconds are spent. Vague prompts are why generated video comes back looking like nothing in particular — this builds the specific version for you, and you can render it here without leaving the page.
Sora 2 Prompt Generator
Two briefs, and what came back
Both clips below were generated on Kavel from prompts written in this structure — they are our runs, not stock footage. Every output is AI-generated, so the same brief will not return the same frames twice; read them as evidence that structure survives the trip, not as a template that reproduces exactly.
A kitchen sprint, specified down to the lens
The brief named the camera package (65mm anamorphic), the aspect ratio, the grade — teal in the shadows, warm amber in the highlights — and then split fifteen seconds into a handful of shots with a stated move for each. What came back holds the grade across cuts and keeps the chef readable through the fastest beat. Naming the optics and the colour is what stops a model from averaging toward generic footage.
Open the video studioA snow sequence, timed shot by shot
The second brief pushed on timing rather than optics: 2.35:1, twenty-four frames a second, and a multi-scene action sequence with each scene given its own slice of the fifteen seconds. The clip keeps the frame rate feel and the widescreen framing, and the scene changes land where the brief put them. Where it strains is continuity between scenes — the further apart two beats look, the more the model treats them as separate ideas.
Try your own briefWhat is a Sora 2 prompt generator?
A Sora 2 prompt generator takes the sentence you would normally type — "a chef running through a kitchen" — and expands it into the kind of brief a director would hand a crew: how many shots, what each one frames, where the camera moves, which lens and colour treatment, and how the running time is divided between the beats. The generation model itself has not changed.
What changes is how much of the decision-making you have made before it starts, and that is most of the difference between a clip you keep and a clip you regenerate five times.
The habit transfers: a brief written this way reads the same to any modern video model, which is why the output here is text you can carry rather than a setting buried in an interface. If you would rather start from a finished effect than from a blank line, the dance video generator and Veo 3.1 pages come with the prompt already written.
A brief written this way survives the trip between engines, because it names things every video model understands: how many shots, what each one frames, where the camera moves, and how the running time divides between the beats.
What it will not do is pin the frames. Video models are sampled, so two runs of identical text differ in framing and timing — structure narrows the range rather than fixing the result, which is why picking from a few takes beats fighting one.
Creative engine
From one line to a shot list
Describe the clip in a sentence and the brief comes back filled in — shots, camera moves, lens and grade. Render it here on Veo 3.1 or Seedance 2.0, or copy it out.
How to write one that survives the render
Each step is a choice you make, not a button you press. Make all three and the brief stops being a wish list.
- 01
Say what the camera is doing
"A woman walks through a market" leaves the camera unassigned, so the model invents one, differently each run. Give it a job — locked off, tracking behind her shoulder, craning up as she passes — and the shot stops drifting between attempts.
- 02
Spend the seconds on purpose
A clip is short. Decide out loud how many beats fit and what each one is for: three seconds establishing, six on the action, three to land it. Models follow an explicit division of time far more reliably than they infer one from a description.
- 03
Name the look, not the mood
"Cinematic" means nothing in particular. A lens, an aspect ratio, a grade — 65mm anamorphic, 2.35:1, teal shadows and amber highlights — are instructions. Swap one of them and you can see exactly what moved, which is what makes the next revision fast.
Where a written brief pays for itself
Three people who stop regenerating and start revising once the prompt has a shape.
The creator burning credits on retries
Ten runs of the same vague sentence produce ten unrelated clips, and none of them can be improved because there is nothing specific to change. A brief makes the next run a revision instead of another roll of the dice.
The marketer who needs the same look twice
A campaign needs six clips that belong together. Fixing the lens, the ratio and the grade in writing is what makes them siblings — the subject changes between briefs, the treatment does not.
The team comparing engines
Judging two models on two different prompts tells you nothing. Hold the brief still and swap the engine and the comparison finally means something — run it here on Seedance 2.0, or take the same text elsewhere.
Questions about prompting for video
What this tool does, what it does not, and what to expect from the text it hands you.
No, and it does not claim to. What it produces is a written brief. Kavel renders it on the video engines wired into this site — Veo 3.1 and Seedance 2.0 — and the same text can be pasted into whichever generator you use elsewhere. If a page ever tells you it is running a model it is not, treat everything else on it as unverified too.
No. Video models are sampled, so two runs of identical text differ in framing, timing and detail. Structure narrows the range — the shot count, the camera moves and the grade hold up across runs — but it does not pin the frames. Plan on picking from a few takes rather than expecting one exact result.
Long enough to answer the questions the model would otherwise answer for you: how many shots, what each frames, how the camera moves, how the seconds divide, what the look is. Past that, extra sentences mostly add contradictions, and a contradiction is the one thing a model resolves at random.
Yes, with one change of emphasis. When you start from a photo, the frame and the subject are already decided, so the brief should spend its words on motion and timing rather than on composition — what moves, in what order, over how many seconds.
Writing and copying a brief costs nothing. Rendering it spends credits, and the estimate for the exact model, resolution and duration you picked is shown in the generator before you start; a failed run is never charged. Free credits come with the account, which is enough to see a brief become footage before you decide anything.
Kavel Pricing — Credits for AI Image and Video Generation
Kavel pricing — pay only for what you generate
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
Starter
What's included
- 33,500 credits per year
- Works with Nano Banana images and Seedance and Veo video
- Cost preview before you generate
- Generation history
Secure checkout by Stripe. Card details never touch our servers.
Creator
What's included
- 89,700 credits per year
- Access to Seedance 2.0, Fast, and Mini
- Nano Banana 2 generation and editing
- Priority credit top-ups
Secure checkout by Stripe. Card details never touch our servers.
Studio
What's included
- 166,800 credits per year
- Built for high-volume Kavel generation
- Plenty of room for 4K rendering
- Available across the supported models
Secure checkout by Stripe. Card details never touch our servers.
Credit packs
One-time top-ups — buy extra credits any time you run low.
Credit Pack 1500
What's included
- 1,500 credits
- No ongoing subscription
- Available across the supported models
Secure checkout by Stripe. Card details never touch our servers.
Credit Pack 3600
What's included
- 3,600 credits
- No ongoing subscription
- Available across the supported models
- Use for both video and image generation
Secure checkout by Stripe. Card details never touch our servers.
Credit Pack 9000
What's included
- 9,000 credits
- No ongoing subscription
- Handy for Seedance 2.0 final renders
- Valid for 365 days
Secure checkout by Stripe. Card details never touch our servers.
Powered by
One credit balance for every Kavel image and video model
Use one balance across the supported image and video models (Nano Banana 2 / Nano Banana Pro / Seedance 2 / Veo 3.1), and check the credit cost before each request.
Video models
Image models
Model credit guide
What each model costs in credits
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. Mini is cheaper still. |
| Video | Veo 3.1 | Billed per video (Lite). About 34 credits at 1080p, about 23 at 720p. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. About 8 credits per image. Start here to test. |
Write the brief, then watch it render
Start from one line about the clip you want. The structure comes back filled in, and the studio is on the same page.