Vertical Drama Script Template: Structure, Example, and How to Write One
By Character App Team · · 10 min read
A vertical drama lives or dies in its first three seconds — and that starts on the page. This vertical drama script template gives you the episode skeleton, a worked 60-second pilot scene, and the formatting an AI Director can actually shoot, so your script becomes a rendered episode.
What makes a vertical drama script different
A vertical drama script is not a shrunken screenplay. A feature breathes; a vertical episode sprints. You have roughly 60 to 90 seconds to land one emotional beat and one hook, framed 9:16 for a phone held one-handed inside a swipe feed. Learning how to write a vertical drama script starts with accepting three constraints that rewrite everything you know about pacing, exposition, and staging.
Pace: every second earns its place. There is no room for a slow build. A viewer decides in the time it takes a thumb to twitch whether to keep watching or swipe on, so the script has to move from the first frame.
Hooks: the script is engineered around cliffhangers, not acts. A season runs 60 to 100 episodes, and each one is a chapter that has to end mid-tension so the next swipe feels mandatory. You are not writing a story with a cliffhanger at the end — you are writing a chain of cliffhangers with story in between.
Staging: you write for a vertical frame. That means close, centered compositions — faces, hands, reactions — not wide establishing shots. A two-shot that works in 16:9 loses both actors’ eyes on a phone. Write action that reads in a tall, narrow window, and your script will translate to screen instead of fighting it.
The 3-second hook rule
The single most important rule in the format: earn the next three seconds, three seconds at a time. The cold open has to hit before a viewer’s thumb reaches escape velocity, and the fastest way to lose them is to script the walk-up to the moment instead of the moment itself.
Do not write "Maya arrives at the office and sits down at her desk." Write "Maya freezes in the doorway — her husband is having lunch with the woman who fired her." Open on the reveal, the betrayal caught in the act, the question the viewer needs answered, or the line of dialogue that reframes everything.
A reliable test: if your first line of action could be cut without weakening the hook, cut it. Keep cutting until the episode opens on tension. Everything a viewer needs to understand the stakes can be delivered in the ten seconds after the hook, not before it.
The vertical drama script template (copy and reuse)
Here is the episode skeleton. Copy the text below, fill each bracket, and you have a shootable script. It runs five beats across roughly a minute — the structure most vertical hits reuse episode after episode. The format at large quotes 60 to 90 seconds per episode, but this skeleton targets the tight end, around 60 seconds: a leaner cut hooks harder, and it is also what an AI Director renders in a single Story, so the script and the render line up. Nothing here is gated; it is plain text you can paste into any editor.
Treat the timecodes as targets, not law. The only two non-negotiables are the cold open and the cliffhanger. Everything between them exists to set up and deliver the turn.
- EPISODE [#] — [TITLE] · Runtime target: ~60s (industry range 60–90s) · Logline: [one sentence that is itself a hook]
- BEAT 1 — COLD OPEN (0:00–0:03): open mid-conflict or on a reveal. No setup, no establishing shot.
- BEAT 2 — SETUP (0:03–0:15): one location. Establish who wants what and what stands in the way.
- BEAT 3 — THE TURN (0:15–0:40): the emotional pivot — a secret lands, a line is crossed, a status flips.
- BEAT 4 — ESCALATION (0:40–0:55): consequences hit. Raise the stakes for the character we are tracking.
- BEAT 5 — CLIFFHANGER (0:55–0:60): cut on the question. End one beat before the answer.
A worked example: a 60-second pilot scene
Abstract structure only gets you so far, so here is the skeleton filled in for the pilot of a secret-identity revenge drama — a genre the format rewards because familiar conventions let you move fast. The logline: the new assistant everyone underestimates is the founder’s daughter, back to take the company that erased her family.
Notice what each beat does. It stays in one location, makes exactly one emotional move, and hands off to the next beat with a rising question. The episode ends before the confrontation resolves, so the next chapter opens on the answer — which is itself the setup for a new hook.
- COLD OPEN (0:00–0:03): Nora stands in a gleaming lobby in a cheap suit. A guard blocks her. "Deliveries are around back." She smiles. "I’m the new assistant." Behind her hangs a portrait of the man who ruined her father.
- SETUP (0:03–0:15): The floor writes her off. Her boss, Victor, does not look up. "Coffee. Two sugars. Don’t speak unless spoken to."
- THE TURN (0:15–0:40): Nora sets down the coffee — and a folder. Inside: the fraud paperwork that destroyed her family, signed in Victor’s hand. He goes very still.
- ESCALATION (0:40–0:55): "Who are you?" Nora finally meets his eyes. "You signed my father’s name on that. I’ve waited eleven years to hand it back."
- CLIFFHANGER (0:55–0:60): Victor reaches for the phone. Nora slides one more page across the desk. "Call anyone, and this goes to the board first." Cut to black on his hand hovering over the receiver.
Cliffhanger engineering across an episode arc
One great cliffhanger is luck. A season of them is architecture, and you plan the architecture before you write the scenes.
The reliable pattern is the nested question: every episode answers the previous cliffhanger in its cold open, then opens a bigger one by the last beat. Answer small, ask big. That rhythm is what turns a casual swipe into a binge — the viewer always has an open loop pulling them forward.
Map the arc on a single page first. Write the question each episode ends on, top to bottom. If two consecutive episodes close on the same kind of beat — two reveals, two arguments — vary one of them. Momentum comes from escalation and rhythm, not from any single shocking moment.
Save your biggest reversal for a mid-season turn — around episode 15 to 20 of a 60-episode run — and plant it early. This is where character consistency stops being a rendering detail and becomes a storytelling tool. You can seed a face, an object, or a relationship in episode one and pay it off twenty episodes later, but only if the audience recognizes it when it returns. If your lead looks like a different person between chapters, the payoff never lands.
Formatting scene actions an AI Director can shoot
If you plan to render your script rather than film it, a few formatting habits decide whether a scene shoots clean or fights you. An agentic AI Director renders a shot, not a montage, so write to that reality.
Write action as one continuous physical moment per scene. "She opens the letter, gasps, then runs to the car" is three shots — split it into three scenes. Externalize emotion, too: a model cannot render "she feels betrayed," but it can render "her jaw tightens and she sets the glass down too carefully." Write what the camera sees, not what the character feels.
Name your shot. Vertical framing wants close, so say so — shot size (tight close-up, medium) and movement (slow push-in, locked). Keep dialogue to one or two short lines and hand each one to a specific character on screen. Use this per-scene format as your unit of work:
- SCENE [#]
- LOCATION: [one place]
- CAST: [which characters are on screen — name them]
- ACTION: [one continuous physical moment, present tense, what the camera sees]
- CAMERA: [shot size + movement — e.g., tight close-up, slow push-in]
- DIALOGUE: [1–2 short lines; subtext over exposition]
From script to screen: direct it as a Story
A script is a plan; the format is only finished when it is on screen. This is the step almost no writing guide covers — turning the draft into a rendered episode. In a making tool you do not hand the script to a crew, you direct it yourself, scene by scene.
In Character App, that plan becomes a Story — a multi-scene episode that renders up to about 60 seconds, which is exactly why the template above targets that length: one episode maps to one Story. First build your Cast: create each character from a single photo or a text prompt and lock it, so the same face, wardrobe, and style render in every scene and every episode. That locked reference is the fix for the drift that makes standard AI clips unusable for a series.
Then direct the Story scene by scene using the format above — write the action, set the camera direction, and assign the Cast on screen for each shot. Reorder beats, rewrite a line, reshoot a scene that is not landing. This is what "interactive" means here: you direct the production, beat by beat. It is not a branching, choose-your-own-path experience for viewers — the finished output is a single linear vertical video you export. There is no file to import; you write or paste each scene directly into the Director as you go. For the full directing walkthrough, see /blog/how-to-make-a-vertical-drama-with-ai, and the /vertical-drama-maker page shows the episodic workflow from the tool side.
Render economically. Draft cheap with a fast model to check pacing and blocking, then upgrade only your hero beats — the cold open and the cliffhanger — to a cinematic model like Seedance 2, which we run in production for clips up to ~15 seconds. You choose the model per scene, so premium Credits go only where they show, and every shot renders 9:16, framed for TikTok, Reels, and Shorts.
On cost: plans start at $19.99/mo, and a free tier covers daily Images so you can build and lock your Cast before spending anything on video. Video generation itself is not free, and every paid action shows its Credits price before you confirm — see /pricing for the full roster.
What vertical-drama writing pays right now
The reason to get fluent in this format now is simple: demand for scripts is outrunning supply. The studios and apps behind the boom are producing at industrial volume and paying writers to feed the pipeline.
As of 2026, trade and industry coverage of the micro-drama market has reported freelance vertical-drama script rates in the rough range of $40 to $100 per finished minute, with the top of that band going to writers who hit the format’s beats reliably and turn scripts around fast. Rates vary widely by studio, region, and experience, so treat any single figure as a snapshot rather than a standard.
Note the unit: per finished minute, not per page. That is the format’s economic logic in one phrase. A 60-episode season at roughly a minute an episode is about an hour of finished runtime, and the writer who can produce hook-per-episode structure at that scale is the one who gets hired again.
Whether you are writing to sell a script or writing to produce your own series, the template above is the same. The only thing that changes is what you do with the final draft — hand it to a buyer, or direct it yourself into a rendered pilot.
Frequently asked questions
How long should a vertical drama script be?
The format runs 60 to 90 seconds per episode; aim for the tight end, around 60 seconds — roughly 120 words of dialogue and action combined. If you plan to render it, target ~60 seconds exactly, because a Story renders up to about 60 seconds and one episode maps to one Story. Seasons typically run 60 to 100 episodes. The discipline is one clear beat and one hook per episode — if you have two turns, you have two episodes.
What is the difference between a micro drama script and a vertical drama script?
In practice, almost none. "Micro drama" and "vertical drama" describe the same 9:16, cliffhanger-driven format, and writers use the terms interchangeably. "Micro" emphasizes the very short episode length; "vertical" emphasizes the phone-native aspect ratio. The same script structure serves both.
Do I need screenwriting software to use this template?
No. The template here is plain text you can write in any editor or notes app. If you plan to render it, you do not import a script file at all — you write each scene’s action and camera direction directly into a Story as you direct, so the format doubles as your shooting checklist.
Can I turn the script into an actual video?
Yes — that is the point of directing it as a Story. Your written scenes become rendered 9:16 shots with a consistent Cast: draft cheaply to check pacing, then render your hero beats with a cinematic model. The output is a single linear vertical video you export and post to TikTok, Reels, or Shorts.