A café owner notices a single wilted basil plant on the counter at closing time. She lifts the pot, frowns at the dry soil, carries it to the sink, runs water until it pools at the base, sets it under the window light, and watches a single leaf straighten. Twenty seconds, six shots, one question answered. That micro-narrative fits a phone screen because every decision—framing, motion direction, cut timing—was made for a narrow vertical canvas before any footage was generated.
Storyboarding a short vertical AI video means planning what the viewer sees, when they see it, and why each shot changes, all within the constraints of a format designed for one-handed mobile viewing. This guide walks through the practical choices that turn a story idea into a shot-by-shot blueprint optimized for 9:16 delivery.
Understanding the Vertical Canvas
A vertical frame is roughly one-third the width of a traditional landscape composition. That narrow window changes how you stage information. Platform interface elements—profile icons, like buttons, comment threads, progress bars—occupy unpredictable portions of the screen depending on the app, device, and user interaction state. Many viewers scroll with sound off.
Test your key visual elements at actual phone size within a conservative central zone. Faces, essential actions, and readable text should sit comfortably in the center of the frame, away from top and bottom edges where UI overlays or device notches may intrude. Avoid assuming universal safe areas; instead, compose for the smallest common viewing rectangle and verify on a physical device.
Crafting an Honest Visual Hook
The opening image should raise a specific story question that the remaining shots will answer. A close-up of the café owner's puzzled expression as she examines the wilted basil creates immediate curiosity: what's wrong, and will she fix it? An honest hook shows a reaction, an unexpected object, a visible state change, or a contradiction between expectation and reality.
Decorative establishing shots or unrelated spectacle can dilute the payoff. If the first frame doesn't connect directly to the narrative problem, viewers may swipe before the story begins. Choose an opening that makes the viewer want to see what happens next, not one that merely looks polished.
Staging Information in Vertical Space
Arrange visual information across three vertical zones—upper, middle, lower—and use depth to separate foreground subjects from background context. Upward motion (a hand lifting a pot toward the camera) and downward motion (water pouring into a sink) naturally fit the tall frame. Toward-camera movement (the owner walking forward with the plant) uses depth rather than width.
Horizontal travel presents a challenge. A character walking left to right across the frame will quickly exit a narrow composition. Solutions include tracking the camera to follow lateral motion, cutting to a new angle that reorients the action vertically, or staging the movement as a series of approach-and-pause beats that keep the subject centered. The basil-rescue story avoids sustained horizontal chase by staging each action as a stationary task or a move toward the lens.
Defining Shots with Purpose
Each storyboard panel represents one meaningful piece of new information for the viewer. A shot changes when the action changes, the camera needs a better vantage point, or the story requires a shift in emotional tone. In the basil example:
| Shot | Action | Camera | Time | Sound |
|---|---|---|---|---|
| 1 | Owner notices wilted plant, frowns | Close-up, static | 3s | Ambient café hum |
| 2 | Lifts pot, inspects dry soil | Medium, slight tilt down | 3s | Ceramic scrape |
| 3 | Carries plant to sink | Medium, follows subject forward | 4s | Footsteps, water running starts |
| 4 | Water pours into pot, pools at base | Close-up, overhead angle | 4s | Water flow, ceramic clink |
| 5 | Sets pot on windowsill, sunlight hits leaves | Medium, static | 3s | Pot settles, ambient fade |
| 6 | Single leaf straightens | Extreme close-up leaf | 3s | Subtle rustle, satisfied exhale (off-screen) |
This six-shot structure is illustrative, not a platform rule. The timing reflects the story's natural rhythm; actual export specifications and duration limits vary by platform and change over time, so check current guidelines before finalizing your project.
Each cut is motivated: we move closer to see the soil, follow the owner's movement, shift to an overhead angle to show the water action clearly, pull back to establish the new location, then zoom in for the payoff detail. Unmotivated cuts—changing angle without a storytelling reason—confuse viewers and waste generated footage.
Planning Audio and Captions
Storyboard columns should include dialogue, voiceover, sound effects, music cues, and intentional silence. In the basil story, the ambient café hum grounds the opening, the water-running sound bridges shots three and four, and the final exhale punctuates the resolution. Board these elements separately so you can source or generate each layer with appropriate rights and quality.
Captions are added during postproduction, not baked into generated frames. Plan for editable text overlays with readable line length, high contrast against the background, and timing synchronized to speech or key actions. Accurate captions serve viewers who watch without sound and improve accessibility; they are not optional decoration.
Preparing Reference Frames
Before generating motion, create rough reference frames for each storyboard panel. These can be sketches, photo composites, or simple digital mockups that establish composition, subject placement, lighting direction, and color palette. Testing static frames on a phone screen reveals whether faces are large enough, actions are clear, and text is legible.
Reuse approved character designs, locations, and props across shots to maintain visual consistency. If the café owner's apron is green in shot one, it should remain green in shot six unless the story justifies a costume change. Recurring assets reduce the cognitive load on viewers and make your limited shot budget more efficient.
Adapting Horizontal Relationships
Many story ideas begin as horizontal scenes: two characters facing each other, a wide landscape, a side-view action sequence. Translating these relationships into vertical composition requires coverage strategy, not blind cropping.
An over-the-shoulder shot can preserve a conversation's spatial logic within a narrow frame by showing one speaker's back and the other's face. A detail shot followed by a tilt or pan can reveal a tall object that wouldn't fit in a single static frame. Cutting between close-ups of each participant's reaction maintains the emotional exchange without requiring both figures in the same shot.
Prepare portrait-oriented source frames whenever possible. If you must adapt existing horizontal material, identify the essential visual relationship in each moment and choose an angle or framing that preserves that relationship vertically.
Building and Testing an Animatic
An animatic assembles your reference frames in sequence with rough timing and temporary audio. It is a low-polish preview, not a finished product. Export the animatic at 9:16 aspect ratio and watch it on a phone in the environment where your audience will see it—standing in line, scrolling in bed, commuting on a bus.
Check whether the opening image raises a clear question, whether each cut feels motivated, whether the focal action remains readable throughout, and whether the payoff delivers on the initial promise. If the animatic doesn't work as a static sequence with rough sound, motion generation won't fix the underlying structural problems. Revise the storyboard, adjust timing, reframe weak compositions, and test again before committing resources to final footage.
Frequently Asked Questions
What aspect ratio should I use for vertical videos?
9:16 is the common full-screen portrait format for most short-video platforms, but export specifications evolve. Check the current technical guidelines for your target platform before finalizing your project.
How many storyboard panels do I need?
One panel per meaningful shot. A 15- to 30-second narrative typically uses four to eight shots, depending on pacing and complexity. The goal is efficient storytelling, not hitting an arbitrary panel count.
Should I add captions during storyboarding?
Plan caption timing and content in your storyboard, but add the actual editable text overlays during postproduction. This allows you to adjust wording, styling, and synchronization after reviewing the generated footage.
Can I just crop a horizontal video to vertical?
Cropping a wide composition often loses essential action, faces, or spatial relationships. It's more effective to plan vertical coverage from the start or adapt horizontal scenes using detail shots, tilts, and angle changes that preserve the important visual information.
How do I handle camera movement in a narrow frame?
Vertical and depth-based motion (up, down, toward, away) works naturally. For horizontal movement, use camera tracking, motivated angle changes, or staging that keeps the subject centered rather than letting them exit the frame quickly.
Deciding Before Generating
A storyboard is a decision-making tool. It forces you to choose what the viewer sees, in what order, and why each shot exists before you invest time and resources in motion generation. The narrow vertical canvas rewards intentional composition, motivated cuts, and staging that uses height and depth rather than width.
The basil-rescue story works because every shot was planned for a phone screen: the owner's frown fits a close-up, the pot travels toward the camera, the water pours downward, the windowsill action is staged frontally, and the final leaf detail is large enough to read. That clarity comes from storyboarding, not from hoping a horizontal composition will adapt gracefully when cropped.
Once your animatic proves the story on a phone, you're ready to generate footage with confidence. If you're exploring image-to-video workflows where you supply a reference frame and describe the desired motion, tools like image-to-video generation can help you test whether your storyboard frames translate effectively into short animated sequences. The available aspect ratios and motion results will depend on the specific model you choose, so treat any generated output as one input to your editing process rather than a finished deliverable.
Plan the story first. The technology follows.