Quick answer
If your photo was labeled because of metadata — C2PA content credentials or an XMP field like DigitalSourceType — you can remove that record from your own file and post the cleaned copy. Most "AI info" and "Made with AI" labels on edited photos come from here.
If detection comes from the image itself — an invisible watermark such as SynthID, or a platform analysing the pixels — removing metadata changes nothing. Anyone promising to make an image "undetectable" is either overselling or altering the image, which this tool does not do.
The two kinds of AI detection
Metadata in the file
- C2PA content credentials
- XMP / IPTC
DigitalSourceType - AI prompts stored in PNG text
- Editing records naming AI tools
Signals in the pixels
- Invisible watermarks such as SynthID
- Visible watermarks and logos
- A platform's visual classifiers
- Human review and reports
Metadata travels with the file like a label on a box. Pixel signals are part of the picture itself. Removing the label from the box doesn't change what's inside.
Step 1: Check which kind you're dealing with
Drop the file into the AI metadata checker (or the tool at the top of this page). If it finds C2PA credentials or a DigitalSourceType such as compositeWithTrainedAlgorithmicMedia, the metadata is the likely source of your label. If it finds nothing, the label is coming from somewhere metadata removal can't reach — and you should stop there.
Step 2: Remove the metadata from your own file
- Cleaning runs in your browser; the file is never uploaded.
- Only metadata blocks are removed. The pixels are copied byte for byte, so quality, colours and rotation stay the same.
- Nothing new is written into the file — no fake camera data, no forged credentials.
Then upload that exact file. Re-exporting it through another editor can write new markers. The Instagram guide covers posting, stories and already-published posts.
When you shouldn't remove it
Metadata removal is for fixing your own files: an edited photo that got a false "Made with AI" label, or a picture whose location you don't want to share. It is not for:
- presenting AI-generated images as human-made to clients, buyers, contests or publications;
- avoiding an AI disclosure that a platform's rules or the law require — under the EU AI Act, for example, transparency duties can apply to whoever publishes AI-generated or manipulated content;
- stripping credit or copyright information from someone else's work.
Removing a metadata field doesn't remove any of those obligations. See the disclaimer for details.
FAQ
Can you remove AI detection from an image?
You can remove the metadata that records AI involvement, which is what most platform labels read. You cannot remove what is in the pixels: invisible watermarks such as SynthID and the platform's own visual analysis are unaffected by metadata removal.
Will removing metadata make an AI image undetectable?
No. It removes the file's record of how it was made. Pixel-level watermarks remain, and platforms can still analyse the image itself. This tool is meant for your own photos and false-positive labels, not for hiding AI-generated content.
Does it remove SynthID or other watermarks?
No. It does not alter pixels, so visible watermarks and invisible ones such as SynthID stay exactly as they are.
Is it legal to remove AI metadata from an image?
Removing metadata from images you own is routine. It is not acceptable to strip copyright information from other people's work, or to use it to avoid a disclosure that a platform or the law requires. This isn't legal advice.
Is my image uploaded?
No. Checking and cleaning run entirely in your browser. Nothing is sent to a server.
Use on files you own and follow platform terms and applicable law.