What are some good AI video prompt examples?
The best examples are prompts with a finished video next to them, so you can see what each line did. Below are excerpts from five prompts behind videos we published, one per format, with what makes each work; the full prompts and videos are in our prompt library.
1. Phone-POV livestream
From Hate Mail, a 30-second Seedance take. It names the format in the first line of dialogue and ends on a turn.
A spoiled, aristocratic, bossy, saucy rich girl reads her hate mail live at breakfast. One uninterrupted take in continuous time through the front camera of the phone propped still on the table, eye level, the only viewpoint, no cuts, no camera movement. 0-4s: She waves the letter at the lens: "Reading my hate mail. Live. Because I'm that hated." 4-9s: She reads the letter in a mocking voice: "'You are a spoiled brat.'" She shrugs: "Correct. Next." She flicks it over her shoulder.
Why it works: the camera position is fixed in one sentence, every beat has one action and one quoted line, and the joke structure (setup, shrug, dismissal) is written into the timeline rather than left to the model. More on the format in our livestream guide.
2. Tutorial
From Love Potion. Tutorials work because each beat is a physical step the model can show.
4-9s: She crushes a handful of red rose petals in her fist and lets them fall into the cauldron, which flares pink: "Rose petals. Crushed by hand. By a girl of breeding." 9-14s: She adds one drop from a tiny crystal vial; the potion swirls into shimmering mother-of-pearl and steam rises in slow spirals.
3. Solo dance to a song
From Look Up. The song is attached as an audio reference, and the prompt says so.
@Audio1: The song she dances to. It plays from the start; every move lands on its beat and her lips sync to its lyrics. 3-13s: The vocals start and she snaps into the dance with a slow hair flip on the first beat, lip-syncing "look up, look up".
Our AI dance prompt page covers music sync in detail.
4. Three-person walk-in trend
From Own It, made as two chained 15-second halves. Group shots need every person re-described and anchored to a position.
@Image1: The opening frame: the three girls standing far back in the bright tent entrance in a V formation (platinum-blonde Malfoid in the middle, Thea with long straight jet-black hair on her left, Odette with dark curls on her right). @Image2: Thea's face. @Image3: Odette's face. Keep all three identities exactly. 0-13s: All three walk straight toward the camera in perfect formation, one step on every beat.
5. Narrated explainer
From Penny, a low-poly character explaining money across 31 clips. Each clip's video prompt is tiny, because the start frame carries the look and a voice-over carries the words.
She strains and drags the huge gold bar a few steps across the cobblestones, sweat drops popping off her head. Camera follows. Low-poly 3D game animation, stylized, smooth motion. Nobody speaks; mouths closed or smiling. No captions, no text overlays, no new characters appearing.
"Nobody speaks; mouths closed or smiling" stopped mouths moving under the voice-over in our renders.
What the good examples have in common
- A fixed camera position stated once, with "no cuts".
- Time ranges, one action each, dialogue in quotes.
- An avoid list: no subtitles, no logos, no visible phone, no change to the face.
- An ending with a turn, so the clip has a payoff.
- Original characters and generic props. Franchise names and famous faces got our renders blocked.
To build your own in this shape, use the free video prompt builder, or read how to write an AI video prompt.
Where Cast fits
- You want to remix one of these: each video page has a button that opens Cast with the same character and idea filled in.
- Livestream, tutorial and dance clips with the free Malfoid & Friends characters (Malfoid, Thea and Odette).
Where it does not:
- 30-second single takes like the examples above: Cast clips run 5, 10 and 15 seconds.
- Narrated explainers with your own voice-over: Cast does not take uploaded audio.
People also ask
- Where can I find real AI video prompts?
- Our prompt library has the full prompts behind every video we published, each next to its video. Prompts without a finished video are much harder to learn from.
- What does a dance prompt look like?
- An audio reference line ("@Audio1: the song... every move lands on its beat"), a locked camera, and a timeline that follows the song's sections. Example 3 above shows one.
- What does a livestream prompt look like?
- A phone-POV camera line with "one uninterrupted take... no cuts", an opening line that names the stream, quoted lines per beat and a turn at the end. Example 1 above shows one.
- Can I copy these prompts?
- Yes, they are published for that. Swap in your own original character and setting rather than the ones shown.
Related answers
- How do I write a prompt for an AI video?
- What prompt makes a good AI dance video?
- Is there a free AI video prompt generator?
- What should an AI character do in a livestream video?
- What is a good prompt for an AI influencer video?
Part of the Cast answers library (Prompting).
Updated