Fast format, same checks
A 24-hour story passes through the same upload scan as a feed post. Nothing about the shorter shelf life exempts it.
Instagram can read hidden AI signals before a story goes live. Process owned story images and videos before upload when Content Credentials, C2PA data, or AI metadata could create a confusing AI Info label.
Drop an image to inspect its Content Credentials locally.
Meta says AI info can appear when its systems detect industry-standard indicators, or when a creator discloses AI-generated content. Stories disappear in a day, but the file still gets read on the way in.
A 24-hour story passes through the same upload scan as a feed post. Nothing about the shorter shelf life exempts it.
Story templates, upscalers, and background removers attach a signed credential plus an XMP block on export. Those XMP fields often survive a re-save that clears the rest.
If a live story already carries the tag, prepare a fresh copy from the source and share that one instead.
Stable Diffusion and ComfyUI write the workflow into a PNG's text chunks: prompt, model, sampler, seed. Plain text inside the file, and a scanner reads it without effort.
The same file signals drive TikTok's AI-generated media disclosure, Pinterest's GenAI labels, and Threads, so a story asset you clean once travels cleanly across all of them.
accepted image and video formats
maximum file size per run
final story asset at a time
A story edit takes two minutes, and it still passes the same file checks as a feed post. These are the tools we see behind a mislabelled story:
CapCut exports. Story edits leave CapCut with their own metadata attached.
Background removers and upscalers. Both are AI now, and both sign the export.
Generative Fill in Photoshop. One fill on one corner signs the whole file.
Lightroom AI Denoise. A raw staple that writes an AI edit marker into the output.
Canva. Templates and magic tools can pass credentials into the download.
Gemini and Veo edits. These add a SynthID watermark on top of the provenance record.
Speed is what catches people out. Nobody inspects a story the way they’d inspect a campaign asset, so the export goes up untouched and the tag arrives with it.
No setting in the app turns the AI info label off while the markers are still in the file. The label is produced from file contents at upload, so there’s nothing to switch. Guides describing a three-dots menu that hides it are describing a screen that doesn’t exist.
A live story can’t be edited either. Instagram gives you no way to swap the media on a story that’s already published, so the fix is delete it, clean the source file, and post the clean version again.
You lose the views and the replies when you do that. Since a story only runs for 24 hours, most of its reach is gone before you notice the tag. That’s the real argument for checking the file before the first upload rather than after.
Photoshop and Lightroom both let you disable Content Credentials in export settings, before the story asset is generated. In Photoshop it’s the Content Credentials panel; in Lightroom it’s a checkbox in the export dialog. Turn it off once and your story exports stop carrying the record.
Our tool handles the other case: the file exists, the credential is already inside it, and re-exporting from the original isn’t practical.
Work from the final export, not a screenshot of it. We check the file for a C2PA Content Credential, a SynthID watermark, and the AI metadata fields we support, then take them off and re-scan the result so you can see what changed. Resolution, duration, and frame rate stay as they were.
One asset at a time is deliberate. A story set is a sequence of separate files, and each one carries its own history, so each one needs its own pass.
Sequence is the whole trick. Clean the file last, because every re-export can write a fresh credential.
Stickers, text, colour, crop to 9:16. Get all of it done, then export a single copy.
Drop that export above. Credentials, watermark, and supported metadata come off in one pass.
Post the file we give back, and keep the original in case you need its provenance record later.
A published story can’t be relabelled. Instagram doesn’t reprocess media that’s already up, and no file tool reaches into a live story to rewrite what’s stored on Meta’s side. So to remove AI info from an Instagram story after posting, you delete it, clean the source file, and share the clean version again.
Stories expire in 24 hours anyway, which makes this cheaper than it sounds. Highlights are the exception: they keep the labelled story alive, so replace the entry rather than waiting it out. And if the story is generated, or altered enough that Instagram asks you to say so, leave the label alone.
Use the exact image or video you plan to share. This page removes and verifies C2PA Content Credentials in your browser, then passes the cleaned file to the homepage SynthID workflow.
Choose the final story image or video. The browser checks it for embedded C2PA Content Credentials and removable metadata.
Create a C2PA-clean copy and confirm the credential is no longer embedded before continuing.
Open Remove SynthID with the C2PA-clean file selected, and finish in the existing SynthID workflow.
We show what’s in the file: credentials, watermark, metadata fields. You see the cause before anything changes.
Supported signals come off in one pass, on a still or across every frame of a story clip.
A second scan confirms the output came back clean, then you download it. Your file never left the browser.
Use the tool for media you own or are authorised to process when the file history doesn’t describe the visible story.
Quick story edits that go up in front of followers and vanish from the feed a day later.
Launch stories, sale graphics, event assets, and campaign clips prepared before publishing.
Authorised retouched images where an export added AI edit markers that don’t match the final photo.
Meta uses AI info language for signals that can come from creator disclosure or from industry provenance records attached to media.
We process accidental file-level noise on content you own. Keep the disclosure when a story is generated, or materially altered in a way Instagram asks viewers to understand.
Because the file carries a signal Instagram can read: a SynthID watermark, a C2PA Content Credential, generator metadata, or your own disclosure. The scan runs at upload, so a normal-looking story can still be tagged.
Yes. Stories go through the same upload checks as feed posts, and the label can appear even though the story disappears after 24 hours.
It means Instagram found an AI-related signal in the file, or a creator disclosed AI-generated content. It describes the file's history, not everything visible in the story.
There’s no switch inside the app. What you control is the file: take the watermark, Content Credentials, and AI metadata off the export, then share that copy.
Export once you’re finished editing, clean that export last, and share it without re-saving through another tool. Each extra export can write a fresh credential.
Yes. A highlight is a story you already published, so the label stays with it for as long as the highlight is up. Waiting it out won’t work here the way it does with a 24-hour story. Clean the source file, post a fresh story, and add that one to the highlight instead.
PNG, JPG, and WebP for images. MP4, MOV, and WebM for video. Up to 50 MB per run, one final story asset at a time.
No. We handle file-level noise on content you own. If a story is generated or materially altered in a way Instagram asks viewers to understand, keep the disclosure in place.
Open the remover page that matches the platform and format you plan to publish.
Run the final image or video through the tool, then publish that copy to Instagram Stories.
Start with a story file