Resize Video for Social Media

Reframe a clip to 9:16, 1:1, 4:5 or 16:9 at 1080p or 720p, either cropping to fill the frame or padding it with bars. Nothing is uploaded.

๐ŸŒ Espaรฑol

Drop your file here (.mp4, .mov, .m4v, .webm)

๐Ÿ”’ Private by design: your files are processed locally in your browser and never uploaded to any server.

One rule that reproduces every platform preset

Look up the recommended export size for TikTok, then for an Instagram feed post, then for a portrait feed video, and you will find four different numbers that appear unrelated.

They are not. Every one of them is the same idea: take the shorter edge of the frame, make it 1080 pixels, and let the longer edge follow the aspect ratio. Vertical becomes 1080 by 1920. Square becomes 1080 by 1080. Portrait becomes 1080 by 1350. Widescreen puts the 1080 on the height and gives you 1920 by 1080.

That single rule is what this tool applies, which is why there is no lookup table of platform names to maintain and no risk of one preset drifting out of date while the others stay correct. Choosing 720p applies the same rule with a smaller short edge, giving 720 by 1280, 720 by 720, 720 by 900 and 1280 by 720.

Reframing a landscape clip for a vertical feed

  1. Drop your video on the box or use Choose a file. MP4, MOV, M4V and WebM are accepted.
  2. Set Resize to. The default is vertical, which is what TikTok, Reels and Shorts expect.
  3. Set Resolution. 1080p is the default; 720p produces a noticeably smaller file.
  4. Choose a Fit mode. Crop to fill trims the edges; pad with bars keeps everything.
  5. Click Resize Video for Social Media.
  6. Download the result. The name records what you asked for, so a vertical crop of clip.mov comes back as clip-resize-9x16-fill.mp4.
  7. Watch it before posting, particularly if you cropped from a wide source.

The filename encoding the settings is more useful than it sounds once you have exported the same clip three ways to compare them.

Crop or pad, and why there is no third choice

A 16:9 frame and a 9:16 frame have nothing in common except their area. You can keep the whole picture or you can fill the whole frame. You cannot do both, and any tool that claims otherwise is doing one of the two and not telling you.

Crop to fill scales the source up until both dimensions cover the target, then takes the centre. Nothing is left empty and something is always lost from the long axis. Going from widescreen to vertical, the loss is severe, because you are keeping a narrow strip out of a wide frame.

Pad with bars does the opposite: it scales the source down until both dimensions fit inside the target, then centres it and fills the remainder with black. Nothing is lost and the picture is smaller within the frame.

For a talking head shot in landscape and destined for a vertical feed, cropping usually wins, because the subject is near the middle and bars waste two thirds of the screen. For a screen recording or anything with text near the edges, padding is the only safe answer.

The upscaling question

Both modes scale in either direction. There is no check that refuses to enlarge a small source, which means a 480p clip asked for 1080p output will be stretched to fill 1080 by 1920.

That does not add detail. It produces a soft 1080p file that is several times larger than the original and no sharper. When your source is smaller than the target, choosing 720p is almost always the better trade: the result looks the same or better on a phone, and the file is a fraction of the size.

To check what your source actually is before deciding, Aspect Ratio Calculator will work out the proportions and the matching dimensions.

Before and after

Trim before you reframe. Trim Video removes the parts you are not going to use, which saves encoding frames destined for the bin. If you need a specific region rather than the centre, Crop Video lets you choose it rather than assuming the middle.

Afterwards, Compress Video will bring the file under an upload limit, and MP4 to GIF turns a short reframed clip into a loop. If the clip is sideways to begin with, Rotate Video should come first. The rest is on the video tools hub.

Frequently asked questions

What exact pixel size do I get?

The resolution you pick is applied to the shorter edge and the other edge follows the ratio. So vertical at 1080p is 1080 by 1920, square is 1080 by 1080, portrait is 1080 by 1350, and widescreen is 1920 by 1080. At 720p those become 720 by 1280, 720 by 720, 720 by 900 and 1280 by 720. Those are the sizes the platforms themselves document.

Which fit mode should I choose?

Crop to fill for anything where the subject sits near the middle, because a full bleed frame is what a vertical feed rewards and bars look like a mistake. Pad with bars when nothing may be lost, such as a screen recording, a slide, or footage with text near the edges. There is no third option that keeps everything and fills the frame; that is a geometric impossibility.

How much gets cut off when I crop?

However much the two ratios differ by. Turning a 16:9 landscape clip into a 9:16 vertical one keeps a strip roughly a third of the original width, so most of both sides is gone. That is the most extreme common case and it is worth previewing before you publish. Going from 16:9 to 1:1 or 4:5 loses much less.

Will a small video be upscaled?

Yes. Both modes scale to reach the target frame, in either direction, so a 480p source asked for 1080p output is enlarged. Nothing is invented by doing that, so the result is a softer 1080p rather than a genuinely sharper one. If your source is small, picking 720p usually looks better and produces a much smaller file.

Why is the output always MP4?

Because reframing requires a pixel filter, so the clip has to be re-encoded whatever happens, and once that is unavoidable there is no cost to standardising on the most broadly playable container. Video is encoded with H.264 at a quality-leaning setting, and the file is arranged so playback can start before the download finishes.

Is the soundtrack re-encoded as well?

It is re-encoded as AAC at 128 kbps when the clip has an audio stream, and simply omitted when it does not. The input is probed first to find out which case applies, because asking the encoder to map an audio stream that is not there fails rather than being quietly ignored.

What are the bars filled with?

Solid black, centred, with the scaled picture placed in the middle. That is the neutral choice and the one that reads as intentional letterboxing rather than as a rendering error. There is no option for a blurred copy of the frame behind the video, which is a different effect and a much heavier one to compute.

Related tools