Video size caps: TikTok 4 GB, Pinterest 2, LinkedIn 500 MB, X 8

The ceilings are 4 GB on TikTok's API, 2 GB for Pinterest video ads, 500 MB for LinkedIn standard video and 8 GB for a default X account, so LinkedIn is the one a long Sume render will hit first. Each number comes from the platform's own page. LinkedIn's Videos API also says the upload initializer accepts up to 5 GB, and the page lists the two figures without reconciling them, so plan for 500 MB.

Length and bytes interact

YouTube's own encoding page recommends 8 Mbps for 1080p at standard frame rates. Taking that as an example bitrate, one minute is 60 MB of video (8 Mb/s times 60 s divided by 8). Then 500 MB lasts about 8.3 minutes, and a full 30-minute video would be 1.8 GB, more than LinkedIn allows. This is arithmetic on an assumed bitrate, not a statement about what Sume produces.

Sume does not publish a bitrate or size for a Timeline 1.0 render, and it rejects codec, crf and ffmpeg_args fields, so you cannot trim bytes by lowering quality. Your levers are length, output size (256 to 2160 per edge) and frame rate.

A size check to run on the result

Download the job's video_url, read the byte length and compare it with the cap of the platform you are sending to. This script does the comparison for the four caps above, using decimal gigabytes and megabytes as an approximation.

import sys

CAPS = {"tiktok": 4_000_000_000, "pinterest": 2_000_000_000,
        "linkedin": 500_000_000, "x": 8_000_000_000}

size = int(sys.argv[1])  # bytes of the finished MP4
for name, cap in CAPS.items():
    print(name, "ok" if size <= cap else "too large")