Stop writing shot lists: a better way to prompt Seedance
Numbered shot lists make Seedance video feel stiff. Learn to convert a shot list into a story-driven Vibe Creating prompt, with a before-and-after example.

There's a reflex a lot of creators have when they sit down to prompt an AI video model: write a shot list. Shot 1, 85mm, slow push-in. Shot 2, close-up, shallow depth of field. Shot 3, cool tones, low saturation. It feels professional and precise — and with Seedance it often produces video that's strangely stiff, like a slideshow of correct-but-lifeless frames.
The fix isn't to write less. It's to write for feeling instead of control. This is the heart of the Vibe Creating approach, and the good news is that converting a shot list into a story-driven prompt is a repeatable move you can learn in a few minutes. Seedance is available on OmniArt, so you can test the difference on your own prompts as you go.
Why shot lists under-deliver
A numbered shot list optimizes for the wrong thing. It tells the model how to operate the camera — lens, move, grade — but says almost nothing about what the scene is about. A capable model will dutifully render each literal instruction and still miss the throughline that makes the shots feel like one piece.
The result is a sequence that's technically on-spec but emotionally flat: the cuts don't build, the motion doesn't mean anything, and the whole thing reads as a list of frames rather than a moment. You over-specified the execution and under-specified the intent.
A before-and-after
Here's the same idea written both ways. The first is a classic three-shot list. The second hands the model the story — a childhood memory that folds into adulthood — and lets it direct.
Regular prompt: "Shot 1: a 7-year-old boy runs in the yard of an old house. A white plastic bag drifts in the sky; he looks up and chases it. Dusk, warm high-saturation light, Tyndall rays. Shot 2: close-up, shallow depth of field — an old woman by the wooden gate, smiling, looking off-frame. Shot 3: same yard, now overcast, cool tones, low saturation; a 30-year-old man stands dead center, hands in pockets, no other motion."
Vibe Creating: "A remembered midsummer afternoon, the old courtyard wrapped in warm amber. A carefree boy races around the yard, chasing a transparent plastic bag the wind keeps lifting just out of reach. By the wooden gate, his grandmother watches him with tender affection. Then time folds over — the boy who chased that bag is now a tired adult, standing in the very same spot. But the gate is empty, the yard gone to seed. The same wind rolls through and lifts the hem of his coat."
Regular prompt
Vibe Creating
Notice the second prompt didn't add camera parameters. It added meaning — the emotional logic that connects the shots — and the model handled the rest.
Note
The example clips in this article are from ByteDance's Seedance "Vibe Creating" practice handbook. They illustrate the method; the same principles apply across the directed video models on OmniArt.
Convert your shot list in four moves
You don't have to start from a blank page. If you already have a shot list, run it through this:
- Strip the technical control. Delete focal lengths, shot numbers, color temperatures, depth-of-field and exposure notes, and rig jargon. None of it describes the scene.
- Find the throughline. Ask what the sequence is actually about — the feeling or the turn at its center. "A man remembers, and the memory is gone" is a throughline. "Three shots of a yard" is not.
- Name the anchor, one action, one mood. Who or what are we watching, what are they doing, and how should it feel? That's the spine of every shot.
- Keep camera intent as a result, not a setting. If a move matters, describe what it should feel like — "the camera drifts closer, and the empty gate sinks in" — instead of "slow push-in, 0.7x."
The output is usually shorter than the shot list and far more alive.
What you keep
Converting a shot list is about removing noise, not meaning. Keep anything that's load-bearing for the scene:
- Hard constraints — dialogue, narration, lyrics, music cues, and sound effects. If you wrote them, preserve them word for word.
- Real story beats — a specific action, a turn, a reveal. Those aren't technical control; they're the scene.
- Genuine style direction — "claymation," "cyberpunk cinematic," "ink-wash" tells the model the world, which is different from telling it the lens.
The test is simple: if a line tells the model how to feel the scene, keep it. If it tells the model how to operate the gear, cut it.
When a shot list is still right
Vibe Creating is for shots where feeling matters more than spec. It's not a universal rule. Keep a precise, parameter-level shot list when the job demands it — exact word-by-word lip-sync, UI and feature walkthroughs, or industrial delivery against a locked set of shots. For the full breakdown of where each mode wins, see how to write Seedance prompts: the Vibe Creating method.
Getting started on OmniArt
Take a shot list you've already written and run it through the four moves above, then generate both versions on Seedance in the video workspace and compare. The shift from "operate the camera" to "tell the story" is the single biggest upgrade most people can make to their AI video prompts — and it costs you nothing but a rewrite.
Ready to Create?
Start generating amazing content with AI