Visible watermarks are the obvious kind: a logo or wordmark in a corner, applied after generation. Invisible ones are more interesting — a pattern woven into the pixels or written into the file's metadata, designed to survive re-encoding and to be readable by a detector rather than an eye.
Whether either exists depends on the provider and, on some models, on a parameter you can set. Where a toggle is exposed it governs the visible mark; provenance signals embedded by the provider are generally not something a generation parameter controls.
Platform-side watermarks are a separate thing again. Downloading a clip through a social app usually stamps that app's mark on it, and re-uploading the stamped file is the single most reliable way to have a video demoted by the platform you are posting to.
In practice
- A watermark toggle, where one exists, is about the visible mark on the frame.
- Never re-upload a file that carries another platform's stamp — export from the source instead.
- Metadata and provenance markers can persist after visible cropping.
The mistake to avoid
Cropping a corner to remove a mark. It changes your composition, and it does nothing at all to anything embedded in the pixels or the file.
Related terms
Safety checker
A safety checker is an automated filter that inspects prompts, outputs or both and blocks material a provider does not permit.
Resolution
Resolution is how many pixels an output contains, usually named by its height — 720p, 1080p, 4K — and set before generation rather than after.
Aspect ratio
Aspect ratio is the proportion between an output's width and its height, written as two numbers — 16:9 for landscape, 9:16 for a full phone screen.
Seed
A seed is the number that decides the random starting noise for a generation, so the same seed with the same settings reproduces the same output.
Prompt
A prompt is the written instruction a generative model reads to decide what to make — the one input almost every model requires.
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.