Vertical video uses a 9:16 aspect ratio (1080×1920 or taller) — the native format for mobile-first platforms like TikTok, Instagram Reels, and YouTube Shorts. Unlike cropping horizontal footage to vertical (which loses 70%+ of the frame), native vertical generation composes specifically for the tall frame: subjects centered vertically, action flowing top-to-bottom, and text areas positioned for platform UI overlays (captions, buttons, profile icons).
Seedance 2.0 generates directly in 9:16 (1080×1920). The model composes for the vertical frame from the start — subjects are vertically centered, backgrounds fill the tall canvas, and there's no cropped-from-horizontal awkwardness.
Divide your prompt into temporal segments — beginning, middle, end — within a single generation. Script a hook in the first 2 seconds, a reveal at second 3, and a call-to-action at the end. The model follows the timeline structure sequentially.
Seedance 2.0 is the AI model behind TikTok Symphony, ByteDance's official creative assistant. The same model architecture that powers Symphony's ad generation is available through PonPon, giving you direct access to ByteDance's vertical video capabilities.
Generate clips in the 3–6 second range that matches the typical attention window for short-form social content. Long enough for a complete micro-narrative, short enough to loop seamlessly in feeds.
The model avoids placing critical visual information in areas typically overlaid by platform UI — the bottom 15% (captions and buttons on TikTok) and top 10% (status bar on Reels). Important content stays in the visible center.
Go to PonPon Video and select Seedance 2.0 from the model dropdown.
In the settings panel, choose 9:16 aspect ratio. This sets the output to 1080×1920 vertical format, native for TikTok, Reels, and Shorts.
Structure your prompt with temporal cues: *"[0-2s] Close-up of a coffee cup on a table. [2-4s] Camera pulls back to reveal a cozy café. [4-6s] A woman picks up the cup and smiles at the camera."* This gives the model a clear narrative arc.
Click Generate and preview the result. For the most accurate preview, view it on your phone — desktop browsers don't replicate the mobile feed experience. Check that the hook works in the first 1–2 seconds.
Download the MP4 file. Upload directly to TikTok, Instagram Reels, or YouTube Shorts. No re-encoding or cropping needed — the output is already in the correct format and resolution.
Whether you're a solo creator, an agency, or a brand — every model adapts to how you work.
[0-2s] Blurred close-up of a mysterious object on a pedestal, dramatic side lighting. [2-4s] Camera focuses and pulls back — revealing a sleek wireless earbud case. [4-6s] The case opens, earbuds glow softly. Text overlay space at bottom. 9:16, 6 seconds.
Model: Seedance 2.0 · Aspect: 9:16 · Duration: 6s
A model walks toward the camera on a city sidewalk wearing an oversized beige trench coat. Full-body vertical framing. Wind catches the coat. Natural afternoon light. Shallow depth of field blurs the street behind. 9:16, 5 seconds.
Model: Seedance 2.0 · Aspect: 9:16 · Duration: 5s
[0-2s] Overhead shot of an empty white plate on a marble counter. [2-4s] Hands place sushi rolls one by one in a circle. [4-6s] Final garnish of microgreens, chopsticks placed beside the plate. Bright, even lighting. 9:16, 6 seconds.
Model: Seedance 2.0 · Aspect: 9:16 · Duration: 6s
[0-3s] A person in workout clothes does a single pushup in a gym. Camera at floor level, close. Sweat visible. [3-6s] They stand up, wipe their forehead, and give a thumbs up to camera. Motivational energy, gym ambient lighting. 9:16, 6 seconds.
Model: Seedance 2.0 · Aspect: 9:16 · Duration: 6s
Generate dozens of vertical ad variations per day for A/B testing. Change the hook, the product angle, or the call-to-action while keeping the same brand look. Faster iteration than filming and editing each variant manually.
Create launch teaser videos — a product reveal in 6 seconds with a built-in narrative arc (mystery → reveal → CTA). Timeline prompting ensures the reveal hits at exactly the right moment for maximum engagement.
Script the opening hook in the first 2 seconds to stop the scroll, followed by the value proposition, then a closing CTA. Timeline prompting maps each segment to specific seconds in the clip.
Generate "authentic-looking" vertical videos with handheld camera feel, natural lighting, and casual framing. Useful for brands that want UGC-style content at scale without recruiting creators.
| Seedance 2.0 (9:16 Native) | Other Video Models | |
|---|---|---|
| Vertical support | First-class 9:16 — compositions designed for the vertical frame, not cropped from 16:9 | Most models treat 9:16 as an afterthought — same composition logic as 16:9 rotated |
| Timeline prompting | Built-in temporal segments — script hook/body/CTA within a single prompt | Kling 3.0: single prompt, no timeline structure. Sora 2: storyboard mode but no native timeline syntax |
| Social platform fit | Platform-aware safe zones — avoids UI overlay areas for TikTok, Reels, Shorts | Other models don't account for platform overlays — critical content may be hidden by buttons |
| TikTok integration | Same model architecture as TikTok Symphony — official ByteDance creative AI | No direct TikTok integration — third-party tools built separately |
| Speed | Optimized for short 3–6s clips — fast generation suited to high-volume social content | Kling 3.0: optimized for longer cinematic clips. Veo 3.1: higher quality but slower |
Social feed viewers decide to stay or scroll within 1–2 seconds. Use timeline prompting to place your most visually arresting element at [0-2s] — a reveal, a close-up, or unexpected motion. Don't waste the opening on establishing shots.
TikTok and Reels overlay captions, buttons, and profile info on the bottom of the frame. Avoid placing text, faces, or key visual elements in the lower 15% of your prompt composition — they'll be covered on most platforms.
Think top-to-bottom, not left-to-right. Full-body portraits, overhead food shots, tall architecture, and vertical motion (someone standing up, a waterfall) work naturally in 9:16. Landscapes and wide group shots fight the format.
Platform algorithms favor content that loops and retains attention. A tight 3–6 second clip with a clear arc outperforms a longer clip that loses viewers. If you need more time, create a multi-clip sequence and post as a carousel or chain in Flow.
Join thousands of creators, agencies, and brands who use PonPon every day.
Start with free daily credits. No credit card required.