Why a 30-second phone video is 200MB, and what compressing it actually trades away

A photo might be a few megabytes. A 30-second video from the same phone can easily be ten or twenty times that size, because a video is really a sequence of dozens of full frames per second — the data adds up fast, even before accounting for the fact that phones increasingly record at resolutions and frame rates well beyond what most viewing situations actually need.

Why video compression is a different problem than image compression

Compressing a single photo only has to deal with detail within one frame. Video compression also exploits the fact that consecutive frames are usually very similar to each other — instead of storing every pixel of every frame from scratch, a video codec mostly stores what changed between frames, which is why a mostly-static video (a talking head, a slideshow) compresses far more efficiently than something with constant motion across the whole frame, like fast action or a shaky handheld shot.

What actually shrinks a video file

• Bitrate — how much data is allowed per second of footage; lowering it is the main lever for reducing file size, at the cost of more visible compression artifacts in busy scenes.
• Resolution — a 4K clip destined only for a phone screen carries far more pixels than will ever actually be visible; downscaling to 1080p or lower can cut file size substantially with no visible loss on smaller screens.
• Frame rate — footage that isn't fast action rarely benefits from 60fps over 30fps, and halving the frame rate roughly halves the amount of frame data that needs storing.
• Codec choice — a newer codec like H.265 typically produces a meaningfully smaller file than older H.264 at the same visual quality, though not every device and platform supports it equally well.

Doing this without uploading a large file somewhere

Compress Video re-encodes a video at a lower bitrate and, optionally, resolution, running through WebCodecs — a browser API that gives web pages direct access to the same hardware video encoding your device's native apps use — so the compression happens on your device rather than requiring the file to be uploaded to a server first, which matters given how large video files already are before compression even starts.