Burn an 'AI-generated' disclosure line into a clip with caption cues

Platform labels are one layer, and each platform sets its own rules. A burned-in line is another, and it travels with the file if it is reposted. YouTube's disclosure page says labels appear in the player or in the description. A line in the frame does not replace that label; it adds to it.

Cues, not speech

The video captions API takes cues (or segments) with text, start and end in seconds, and does no speech-to-text with them. That is also the path for silent clips. You can send only one of script_text, words, cues, and segments. A job reserves $0.20, priced for videos of at most 60 seconds.

Request

import json, os, urllib.request

body = {
    "video_url": "https://media.sume.com/artifacts/example/final.mp4",
    "style": "slam",
    "cues": [{"text": "AI-generated scene", "start": 0.0, "end": 3.0}],
}
req = urllib.request.Request(
    "https://api.sume.com/v1/video-captions",
    data=json.dumps(body).encode(),
    headers={
        "Authorization": "Bearer " + os.environ["SUME_API_KEY"],
        "Content-Type": "application/json",
        "Idempotency-Key": "disclosure-line-001",
    },
    method="POST",
)
print(urllib.request.urlopen(req).read().decode())

Mind placement

Check the line at phone size and away from the sticker area of the app. design.placement.anchor_ratio moves a line as a fraction of frame height, for styles that support design.