Snapchat Ghost watermark on AI images: can you remove it?

Snapchat puts a Snap Ghost with sparkles watermark on AI-generated images exported from its own tools, and its support page says removing that watermark violates Snapchat's terms. The same page says not all AI-generated images carry a Context Card or watermark, and images made with non-Snap products may not be labeled as AI at all.

All facts here come from Snapchat Support's Generative AI on Snapchat page (read 2026-10-02). The page does not mention C2PA, so I make no claim about it.

What does Snapchat label?

The page lists several AI indicators: sparkle icons, specific disclaimers, Context Cards and tooltips. The stated aim is that people should know when they are interacting with AI or viewing AI-produced content.

The watermark's purpose, in the page's words, is to provide transparency that an exported image was created with generative AI and is not real or based on real events.

What does it not cover?

The page is explicit that coverage is partial. Not every AI image gets a Context Card or watermark, and images created with non-Snap products may be unlabeled. So an image you generate elsewhere and post to Snapchat is not automatically marked by Snapchat.

That leaves the label to you if you want one. This matters for anyone who generates product images or short clips with an external tool and shares them on the app.

Can I remove or replace the watermark?

The page says not to remove the Ghost watermark, because doing so violates its terms. That applies to Snapchat's own watermark on images from its tools. It is not a statement about your own label on content you made elsewhere.

If you generate with Sume, the output carries no Snapchat watermark, and Sume's docs describe what each tool returns and say nothing about a watermark or embedded provenance record on outputs, so treat the output as unmarked until you have checked the delivered file yourself. To add a visible notice to a clip, Video captions burns authored cues, and Timeline compose can place a prepared still over a video for one shot. Sume does not decide whether a notice is required for your post.

Why would a creator care about the watermark?

A visible mark on exported images affects how a post looks, and some creators will want it gone. The support page leaves no ambiguity that doing so is against the terms, so the choices are to keep the mark, create the image outside Snapchat's tools, or accept the label.

If you create outside Snapchat, remember the page's own warning that such images may not be labeled as AI. That puts the disclosure question on you and on whichever other platform rules apply to the post, such as an ad policy or a local labeling law.

A calm way to handle it is to decide the label per campaign, not per tool: write the one-line notice you want on AI-made assets, burn it in once, and reuse it everywhere. That keeps the story consistent when the same asset appears on several platforms with different native labels.

What I would do

Use this as a working list and adjust it to your own pipeline and counsel's advice.

• Do not crop or paint out Snap's watermark on exported images.
• Add your own short notice to AI content made outside Snapchat if you want it labeled.
• Do not assume Snapchat will flag content made with other tools.
• Read the live support page before a campaign, since policies change.