TikTok Video Size and AI Workflows for Every Platform

If a vertical clip looks razor-sharp on your phone and slightly soft the moment it lands in the TikTok feed, the export is usually the culprit rather than the camera. Every short-form platform re-encodes what you upload, and each one expects a slightly different shape. Get the shape right and the compression has far less to fight with.

The short answer for TikTok: 1080 × 1920 pixels, a 9:16 aspect ratio, MP4 container with H.264 video and AAC audio, 30 or 60 fps, and a bitrate around 8–15 Mbps for 1080p. Generate or shoot at a higher master resolution, then export down to that target. Everything else in this guide is about making that number survive contact with the other platforms you probably post to.

The Short Answer: TikTok's Actual Video Specs

Resolution and aspect ratio

TikTok is a phone-first, full-screen experience, so vertical 9:16 is the native shape. 1080 × 1920 is the upload target. A 2160 × 3840 vertical master is useful for editing headroom, but the platform will compress it back down, so there is no benefit to shipping 4K straight into the app.

Landscape and square uploads technically work, but they get letterboxed or pillarboxed inside a vertical frame. That wastes 30–50% of the visible screen and hurts retention in the first second, which is exactly when the algorithm is deciding whether to show your video to anyone else. If you have a horizontal asset you love, crop it deliberately instead of letting the platform pad it with black bars.

Container, codec, frame rate, bitrate

Stick to MP4 with H.264 for video and AAC for audio at 44.1 or 48 kHz stereo. It is the least surprising combination across mobile uploads and desktop uploads alike.

• Frame rate: 30 fps for talking-head, tutorial, and text-driven content; 60 fps for fast motion, sport, dance, and anything with quick camera moves.
• Bitrate: 8–15 Mbps at 1080p is a reliable range. Going much higher mostly inflates file size before the platform re-encodes it anyway.
• File size: mobile uploads are capped around a few hundred megabytes, while desktop uploads tolerate several gigabytes. If a file refuses to upload, re-export at a lower bitrate before you start deleting shots.

Length and the loop

Many accounts can publish up to ten minutes in-feed, but completion rate drops sharply past the one-minute mark. The practical sweet spot for most creators is 15–34 seconds. More importantly, design the final frame to lead naturally back into the first frame. A clean loop earns rewatches, and rewatches are one of the strongest signals a short-form feed can read.

Why One Export Fits Nothing

Cross-posting a single file is the most common reason a good idea underperforms on three platforms at once. TikTok, Reels, and Shorts all use vertical frames, but their interfaces sit on top of that frame differently, their compression settings differ, and their audience expectations differ.

Destination	Native shape	Master to keep	Notes
TikTok	9:16	1080 × 1920	Full-screen vertical, UI on right and bottom
Instagram Reels	9:16	1080 × 1920	Similar frame, heavier bottom overlay
YouTube Shorts	9:16	1080 × 1920	Title text sits low, longer runtime tolerated
Instagram feed	4:5	1080 × 1350	More screen area than a square post
YouTube long-form	16:9	1920 × 1080	Chapter markers and thumbnails matter more
LinkedIn / X feed	1:1 or 16:9	1080 × 1080	Autoplay muted, subtitles are essential

The takeaway is not that you need six separate productions. You need one vertical master composed so it can be reframed into 4:5 and 1:1 without losing the subject, plus a finishing pass per destination for captions, cover frame, and audio.

Safe Zones, Overlays, and the Crop That Eats Your Text

The most expensive mistake in vertical video is putting information where the interface will cover it. On TikTok, the right edge carries action buttons, the bottom third carries the caption, username, and navigation, and the top strip carries navigation and occasionally a banner.

Treat the usable area as a center band of roughly 1080 × 1400 pixels. Keep faces, product labels, and any baked-in text inside that band, and let backgrounds extend into the edges. If you generate footage with AI, this is a composition instruction, not a post-fix: ask for the subject centered in the middle third with empty space at the top and bottom.

A quick test before publishing: crop the frame to 1:1 and then to 4:5. If the story still reads in both crops, your composition is resilient. If the punchline disappears, the shot is too edge-dependent.

Design Vertical-First Footage That Survives Reframing

Horizontal thinking is hard to unlearn, especially if your reference points are film and television. Short-form vertical rewards a different grammar:

1. One subject, one action per shot. Two competing subjects force the viewer to choose where to look, and they usually choose the scroll.
2. Depth instead of width. Since you cannot spread information across a wide frame, layer foreground, subject, and background to create visual interest inside a narrow column.
3. Movement toward the camera. Pushes, reveals, and forward motion read far better in 9:16 than lateral pans, which crop awkwardly.
4. Text as design, not decoration. Leave deliberate space for captions rather than scattering words across the frame and hoping.
5. A readable first frame. The cover frame is a thumbnail. Design it as one.

If you are generating clips rather than filming them, choose a tool that treats aspect ratio as a first-class setting. Orelon's AI video generator lets you start from a vertical concept instead of cropping a horizontal clip later, which preserves both resolution and composition.

An AI Workflow From Concept to Platform-Ready Exports

Here is a repeatable sequence that works whether you post three times a week or three times a day.

Step 1 — Lock the concept in one sentence

Write the promise of the video in a single line: "A 20-second reveal of how a ceramic mug is glazed." If you cannot compress it into one sentence, the edit will sprawl. This sentence also becomes the hook line for your first frame of text.

Step 2 — Build a shot list of five to eight beats

Short-form videos fail when they try to hold twelve ideas. Five to eight beats is enough for a hook, three to five developments, and a payoff. Assign each beat a duration before you generate anything, because timing constraints force clearer shots.

Step 3 — Generate the vertical master first

Produce every clip in 9:16 at the highest resolution available. Do not generate 16:9 and crop; you lose roughly half your pixels and the framing will feel accidental. If you want a faster start, browse video templates for structures that already respect vertical composition.

Step 4 — Reframe rather than regenerate

Build a 1080 × 1920 master timeline. For 4:5 and 1:1 versions, reposition the crop with keyframes so the subject stays centered. Regenerating for each platform burns time and introduces inconsistency in lighting, wardrobe, and character appearance.

Step 5 — Finish per platform

Export the vertical master once, then produce a 4:5 and a 1:1 variant from the same timeline. Add captions per version, since text placement that clears TikTok's interface may collide with Reels' overlay. Normalize audio to roughly −14 LUFS so nobody reaches for the volume slider.

Prompt Patterns That Keep Compositions Reframe-Proof

Generating footage is only half the job; describing the frame is the other half. Vague prompts produce centered-ish subjects that drift toward the edges, which is exactly what breaks on reframe. Include framing, safe space, and the absence of baked-in text in every prompt.

Vertical 9:16 shot, subject centered in the middle third,
generous negative space at top and bottom, no on-screen text,
soft window light from the left, shallow depth of field,
slow push-in, 4 second duration, cinematic color grade

Three habits make prompts more reliable:

• State the aspect ratio first. It anchors everything that follows.
• Describe where the subject is, not just who they are. "Centered, chest-up, looking slightly off-camera" beats "a person."
• Reserve empty space explicitly. Empty space at the top becomes your headline; empty space at the bottom becomes your captions.

Reusable prompt fragments are worth storing. A prompt library saves you from re-deriving lighting and framing language for every new clip, and consistency in prompts translates directly into consistency on screen.

Platform-Specific Finishing Touches

TikTok. Hook in the first 0.8 seconds, use native captions, and keep the caption text short enough that it does not cover the subject. Trending audio still helps discovery, but only when it fits the pacing.

Instagram Reels. The bottom overlay is taller and the cover frame is more prominent in the profile grid. Export a 4:5 variant for feed posts rather than letting the vertical version get cropped automatically.

YouTube Shorts. Titles carry more weight here because the text sits low and stays visible. Runtime tolerance is higher, so a 60-second Short is not penalized the way a 60-second TikTok often is.

LinkedIn and X. Autoplay is muted and viewers are at a desk. Burn in subtitles, favor square or 16:9, and front-load the claim in the first two seconds.

Mistakes That Quietly Cost You Reach

• Uploading the 4K master. Extra resolution mostly produces a larger file and a longer upload, not a sharper feed video.
• Baking text into the outer edges. The interface covers it, and viewers assume the video was cross-posted carelessly.
• Reusing a horizontal export with blurred bars. The blur reads as filler and shrinks the subject to a postage stamp.
• Over-compressing before upload. Export at 8–15 Mbps, not at the lowest preset your editor offers.
• Ignoring the first frame. A dark, half-loaded opening frame is a scroll trigger.
• Breaking character between cuts. If your video has a recurring presenter or mascot, keep wardrobe, lighting, and framing locked across generated shots.
• Skipping captions. A large share of viewers watch muted, and silent video without text is a dead end.

Pre-publish checklist

1. Resolution is 1080 × 1920 and the aspect ratio is exactly 9:16.
2. All text sits inside the center safe band.
3. The first frame reads clearly as a still image.
4. Captions are legible on a phone at arm's length.
5. Audio peaks are controlled and dialogue is intelligible without headphones.
6. A 1:1 and 4:5 crop have been tested without losing the subject.
7. The final frame loops back to the opening frame.

FAQ

Is 1080 × 1920 still the correct size for TikTok? Yes. It is the native vertical resolution and the safest upload target. Higher-resolution masters are useful for editing, but the platform will compress them to a similar delivery size.

Does aspect ratio affect reach? Not directly, but it affects everything the ranking system measures: how much of the screen your subject fills, whether text is readable, and how long people stay. A padded horizontal video usually loses viewers in the first second.

Can I post the same file to TikTok, Reels, and Shorts? You can, and it will play. It will not perform equally. Each platform covers a different portion of the frame with interface elements, so a caption that clears TikTok's overlay may sit under Reels' buttons. Adjust per destination.

How long should a TikTok video be? For most creators, 15–34 seconds. Longer works when the content is genuinely sequential, such as a recipe or a repair, and when the first three seconds promise a payoff worth waiting for.

Do I need separate files for feed posts? Yes, if you also publish to Instagram or LinkedIn feeds. Export a 4:5 and a 1:1 variant from the same vertical timeline instead of letting the platform crop for you.

What bitrate should I export at? 8–15 Mbps at 1080p is a practical range. Files above that range rarely look better after platform re-encoding, and they slow down mobile uploads.

How do I keep a character consistent across generated shots? Lock the description, wardrobe, lighting direction, and lens language in a reusable prompt block, then generate all shots in the same session with the same framing instructions. Consistency comes from repetition, not from luck.

Should I generate vertical footage or crop horizontal footage? Generate vertical. Cropping a horizontal master throws away roughly half the pixels and produces framing that feels accidental. Reserve horizontal generation for long-form video and desktop-first channels.

Make Your Next Vertical Video With Orelon

TikTok's specs are simple; the discipline is in composition, safe zones, and a finishing pass per platform. Start with a vertical master at 1080 × 1920, keep your subject inside the center band, and reframe rather than regenerate when you publish elsewhere.

Orelon is built for cinematic ideas in motion, so you can generate vertical footage that is framed correctly from the first shot. Try the AI video generator with a five-beat shot list, pull a few structures from the template library, and compare your workflow against other tools on the alternatives page. One master, three crops, and a hook that earns the second second.