A good video prompt names the motion, locks identity, sets the camera and states duration; use brackets as a checklist, then order it the way your vendor asks.
- Vendors differ on order. Seedance 2.5, Runway, Veo 3.1 and Kling 3.0 each publish different prompt structures.
- Brackets are a checklist. They make sure motion, identity, camera and duration are all covered.
- Lock identity in words. Add "no morphing" to the subject, plus a reference photo for high-movement shots.
- Keep the camera separate. Mixed instructions can make the subject move instead of the camera.
- Always state duration. Add "no scene cuts", or some models insert a hard cut mid-clip.
AI video prompting is more demanding than image prompting. You're not describing a single frame — you're describing motion, time, and change. Every element that can go wrong in an image can also drift from frame to frame across a clip. Structure is the most reliable tool you have for keeping it consistent.
This article covers the structure that works: the bracket format, the motion-first principle, identity lock, camera specification, and duration.
An earlier version of this page said the same structure applies to every video model. Checked against the vendors' own documentation, that is not true. They publish different structures — set out side by side on what the vendors actually say — and none of them is this bracket format.
Seedance 2.5 — ByteDance publishes an eight-part formula, and it is not this one. Through Dreamina, their own platform, the stated order is Format → Subject → Action → Environment → Camera → Look → Timing → Audio & Constraints. Format comes first (duration, aspect ratio, shot structure) and action comes third, not first. Their template also binds references with @Image1/@Image2, puts shot size before style language, allows one camera move per shot, budgets roughly 6–8 seconds per beat, and places exclusions last — after a positive Preserve [identity, wardrobe, logo, object geometry] clause. See the full template.
Runway Gen-4/4.5 — no fixed order. Runway: “Prompts do not need to follow a specific structure in most cases to receive quality outputs, but we recommend prompting with full sentences for more control over certain elements and appending additional keywords in cases where more variation is welcomed.” For video it says to start from “only the most essential motion to the scene” and add elements one at a time. An earlier version of this line gave Runway an order ending in motion over time and style, and said a bracket prompt works against it; neither is in Runway’s guides.
Veo 3.1 — Google's published formula is [Cinematography] + [Subject] + [Action] + [Context] + [Style & Ambiance], which opens with the camera, not the motion — the reverse of the motion-first principle below. Veo also has its own audio syntax: effects prefixed SFX:, background prefixed Ambient noise:, and speech written after a colon without quotation marks, in Google’s newest guide (older Google pages say to put it in quotes).1
Kling 3.0 — Kling’s own formula is Subject + Subject Movement + Scene. Its separate negative-prompt field exists only on the legacy API, and Kling recommends writing negatives inside the prompt anyway. An earlier version of this line called it a real negative-prompt field.
What carries across better than the syntax is the thinking: name the motion explicitly, anchor identity so it cannot drift, specify the camera once, and state duration. The vendors disagree on order and packaging.
So what is the bracket format below? It is a working structure that was tested on Seedance 2.0 in June 2026 and produces good output. It is not, and this page should not have implied it was, the structure ByteDance publishes. Use it as a checklist that guarantees you have not forgotten motion, identity, camera or duration — then order the finished prompt the way your model’s vendor asks. Per-vendor templates with sources: what the vendors actually say.
2Why Video Prompts Need More Structure Than Image Prompts
In a single image, a vague subject description produces a generic image. In a video, a vague subject description produces a different-looking person in every frame. The model isn't just guessing the subject once — it's regenerating its interpretation of the subject across every frame, and without explicit anchoring it will drift.
Motion introduces another layer. The model needs to understand what's moving, how it's moving, at what speed, and from what camera angle — all simultaneously. Without clear structure, motion instructions bleed into subject descriptions, camera movements get conflated with subject movements, and the output becomes incoherent.
Video drifts frame by frame. A vague subject can look different in every frame, and loose motion instructions bleed into subject and camera descriptions.
The Bracket Format
The bracket format organises video prompt elements into labelled sections. Labelling each element keeps motion, subject and camera instructions apart, and makes it easy to see that nothing is missing. No vendor documents that a model reads the labels as separate sections.
You don't need all six sections for every prompt. Short lifestyle clips might only need [MOTION], [SUBJECT], [ENV], and [DURATION]. Complex cinematic sequences need all of them. The key is: whatever you include, put it in its own labelled bracket.
Give each element its own labelled bracket. You do not need all six sections; short lifestyle clips may only need motion, subject, environment and duration.
The Motion-First Principle
This page’s bracket format puts motion first, so the one thing a still image cannot show is never left out. Runway’s video guide starts there too: “Begin with a foundational prompt that captures only the most essential motion to the scene.” Other vendors order it differently, and no vendor documents that video models in general weight earlier words more heavily, so treat motion-first as this format’s habit rather than a rule.3
This structure puts motion first so it is never left out. Veo 3.1 and Seedance 2.5 publish different orders, so check your vendor's guide.
Identity Lock — The Most Important Line in Any Video Prompt
Temporal consistency — keeping the subject looking the same across every frame — is the most common failure mode in AI video. Without explicit instructions, the model will subtly change the subject's face, hair, and clothing between frames. Over 8 seconds, this drift is obvious and unusable.
Identity lock is a prompt instruction — not a model setting. You add it explicitly to the [SUBJECT] section:
The phrase "no morphing" is particularly effective — it directly names the failure mode you're preventing. For models that support reference image upload (Higgsfield Studio, ControlNet-based pipelines), use a reference photo in addition to the text lock. Text alone may not be enough for high-movement sequences.
Identity lock is a prompt line, not a setting. Add it to the subject with "no morphing", and use a reference photo too where the model supports one.
Camera Specification
The camera bracket controls how the scene is filmed — not what's in it. Separating camera instructions from subject instructions is essential. When they're mixed, the model often applies movement instructions to the subject rather than the camera.
Key camera movements for AI video:
Tracking dolly — follows the subject laterally. Produces the most natural-looking motion for walking sequences. Dolly push-in — moves toward the subject. Creates intensity and intimacy. Handheld — introduces authentic shake. Good for UGC and documentary aesthetics. Aerial drone — overhead or elevated perspective. Static tripod — zero camera movement. Forces the subject's motion to carry the scene.
You can also name the lens or the focus. Google’s Veo 3.1 guide lists “Lens & focus” among its camera words, with examples: “Shallow depth of field, wide-angle lens, soft focus, macro lens, deep focus.”4 No vendor guide read for this page says how much a focal length such as 35mm or 85mm changes a video model’s result, so treat those numbers as a style hint and check the clip.Reasoning, 13 Oct 2026 — Google’s Veo guide names lens types without saying what each one changes, and no vendor guide read for this page tests focal-length numbers.
Keep camera instructions apart from the subject. Name the movement, such as tracking dolly or static tripod, and add a lens or focus effect, such as shallow depth of field, if you want one.
Duration and Frame Rate
Clip lengths differ by model. Runway: “Gen-4 creates videos in 5 or 10 second durations based on an input image and text prompt you provide.” Its newer Gen-4.5 is more flexible: “Supported durations 2 - 10 seconds”.5 Kling, for Kling 3.0: “The new model generates up to 15 seconds of continuous video, with a flexible duration ranging from 3 to 15 seconds.” Kling 4.0, which Kling says rolled out to all users on 15 Oct 2026, goes up to 30 seconds in one pass, Kling says.6 Dreamina, for Seedance 2.5: “You can create cinematic videos up to 30 seconds in standard mode or extend them to 180 seconds with the beta long-video mode.” Always specify duration, in the prompt or the tool’s duration setting, so the motion you describe fits the clip you get.7
"No scene cuts" is important. Without it, some models will insert a hard cut in the middle of the clip — which destroys the continuity you were trying to build. "Continuous motion throughout" reinforces that the action should flow from start to finish without interruption.
Always state duration. Clip lengths differ by model, so match the motion to the clip; add "no scene cuts" so some models do not insert a hard cut.
A Complete Working Example
Related on this site: Identity Lock Prompt Templates for AI Images and Video · First Frame and Last Frame in AI Video · Sound in AI Video
- 1Google Cloud, Best practices for generating videos, Gemini Enterprise Agent Platform documentation (the older Vertex AI address redirects here), last updated 9 Oct 2026, read at source 13 Oct 2026“To prevent the model from rendering text in the video, use a colon ( : ) after the speaker’s action to denote speech and avoid using quotation marks”; its recommended example is “A woman says: My name is Clara.” ↑
- 2Vendor documentation: Luma and Melies Seedance guides · Runway, Gen-4 Image and Gen-4 Video prompting guides, re-read in a browser 17 Sep 2026 · Google Cloud Veo 3.1 prompting guide · Kling API docs. Checked 25 Aug 2026↑
- 3Runway, Gen-4 Video Prompting Guide, read in a browser 17 Sep 2026. An earlier version of this paragraph said AI video models weight earlier tokens heavily, with no source.↑
- 4Google Cloud, The ultimate prompting guide for Veo 3.1, 16 Oct 2025, read at source 13 Oct 2026“Lens & focus: Shallow depth of field, wide-angle lens, soft focus, macro lens, deep focus.” ↑
- 5Runway, Creating with Gen-4.5, read at source 22 Sep 2026.↑
- 6Kling AI, Kling 4.0 vs 3.0, 30 Sep 2026, read at source 1 Oct 2026“Kling 4.0 extends native generation to 30 seconds in a single pass.” (On 1 Oct the page said early access began 28 Sep and the full model was due in October.) · Kling AI, Kling 4.0 at Work, 27 Oct 2026, read at source 3 Nov 2026: “Since Kling 4.0 rolled out to all users on October 15, creative teams across film, advertising, and animation have been putting these capabilities to work.” ↑
- 7Runway, Gen-4 Video Prompting Guide, read in a browser 17 Sep 2026 · Kling AI, Kling VIDEO 3.0 model user guide, read at source 17 Sep 2026 · Dreamina (ByteDance), Seedance 2.5, read at source 17 Sep 2026. An earlier version of this paragraph said most models support clips up to 10–15 seconds and default to their shortest length, with no source.↑
Free printable AI cheat sheets, and AI news every week.
Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.
SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK