Meta documents industry-standard provenance indicators as one input for labels on images posted to Instagram.
Instagram “AI info” labelPlatform-by-platform guidance
AI labels do not work the same way on every social network
Some platforms read provenance metadata. Others rely on creator disclosures, classifiers, platform tools, reports, or community rules. Start with the network where you plan to post.
Inspect image metadataRuns locally in your browser. Your source image is not uploaded.
Choose a platform
Each guide separates documented platform behavior from the file-level metadata you can inspect locally.
Facebook “AI info” label
Meta documents industry-standard indicators as one way it identifies AI-generated images posted to Facebook.
MetaScrub can inspect still images for Facebook posts, not Facebook video, Reels, or audio.
Facebook “AI info” labelThreads AI information label
Threads
Meta says its industry-standard indicator approach covers AI-generated images posted to Threads.
MetaScrub supports still images attached to Threads posts, not video or audio posts.
Threads AI information labelTikTok “AI-generated” label
TikTok
TikTok says an automatic label may follow use of TikTok AI effects or uploaded content with C2PA Content Credentials.
MetaScrub can inspect TikTok photo-post images or still covers. It does not support MP4 video or audio.
TikTok “AI-generated” labelX “Made with AI” or manipulated-media notice
X (formerly Twitter)
X’s authenticity policy says it may label synthetic or manipulated media; it does not document one removable image tag as the deciding trigger.
MetaScrub supports still images posted to X. It cannot clean video, audio, Community Notes, or post labels.
X “Made with AI” or manipulated-media noticeYouTube altered or synthetic content label
YouTube
YouTube documents creator disclosure, its own generative tools, C2PA metadata, and manual review as possible label paths.
MetaScrub can inspect a JPEG, PNG, or WebP thumbnail or community-post image. It cannot clean a YouTube video or audio track.
YouTube altered or synthetic content labelLinkedIn Content Credentials icon
LinkedIn says cryptographically signed image and video content with C2PA information can display a Content Credentials icon.
MetaScrub supports LinkedIn still images, including post or ad images, but not LinkedIn video.
LinkedIn Content Credentials iconPinterest “AI modified” label
Pinterest says it analyzes Pin metadata following the IPTC standard and also uses classifiers to detect generative AI.
MetaScrub supports still image Pins. It does not support video Pins, classifiers, or ad disclosure settings.
Pinterest “AI modified” labelSnapchat generative-AI Context Card
Snapchat
Snapchat says some images created with its generative-AI features may include a Context Card.
MetaScrub can inspect a still JPEG, PNG, or WebP before sharing. It cannot clean a Snap video, Lens, audio, or in-app AI context.
Snapchat generative-AI Context CardReddit AI tag or disclosure
Reddit’s policy asks people posting permissible AI-generated content to use a tag or another clear indication.
MetaScrub supports still images before upload. It cannot remove subreddit flair, moderator actions, reports, or post text.
Reddit AI tag or disclosureOne cleaner, three important boundaries
Use file cleanup for privacy and provenance control—not as a promise about what a platform will display.
- Still images only: JPEG, PNG, and WebP
- No video, audio, classifier, or pixel-watermark removal
- Current platform disclosure rules still apply
Image metadata is only one possible signal.
MetaScrub can inspect and remove supported metadata inside JPEG, PNG, and WebP files. It cannot change a platform’s label, policy decision, classifier, account context, or published post.
Inspect image metadataRuns locally in your browser. Your source image is not uploaded.