Is this photo AI? Ask the file.

What the check reads

• AI label. The IPTC "digital source type" field. Tools that follow the standard set it to "trained algorithmic media" for an image made by generative AI, or to a composite value when AI was used on part of it.
• Content Credentials. A signed record, also called C2PA, that some cameras, editing apps and AI tools attach. It can say which tool made or changed the picture.
• Camera data. The make and model, the lens and the time the shutter was pressed, as written by the camera or phone.
• Software. The last program that saved the file, such as an editor or a phone's own photo app.
• Location. Whether GPS coordinates are stored in the file. Anyone you send the original file to can read them.

Reading the result

The file carries an AI label

The tool that made the file says AI generated or changed it. That is the clearest answer a file can give. A label is rarely added by mistake, but it can be removed, so its absence on other files means little.

The file carries Content Credentials

A signed record is there. This page shows the tool named in it. To check the signature and the full history, open the file in the official Content Credentials verify tool. Our guide to Content Credentials explains what the record can and cannot prove.

Camera data, no AI label

The file looks like it came from a camera or phone. Free tools can write or copy these fields, so treat them as a hint, not proof.

Nothing at all

This is the most common result. Most social apps and messengers strip this data when a photo is posted or sent, and screenshots never had it. An empty file proves nothing either way.

When the file has nothing to say

No checker can tell you for certain whether a stripped photo is real, and AI detectors that give a percentage are often wrong in both directions. What works better is slower: look for the original source with a reverse image search, check the details that AI still gets wrong, and ask who posted it first. We go through each step in how to tell if a photo is real or AI, and through fake profiles in how to spot AI influencers.

Photos taken in Rawly

Rawly is a photo and video challenge app that does not try to catch AI images. It removes the way in. Every photo is taken with the in-app camera. There are no gallery uploads and no filters. A shared Rawly photo carries a code you can enter at rawly.app/verify to see the capture and the minute it was taken. The code shows when, never who. Someone could still point the camera at a screen, and people can report entries like that. More on this in real photos and AI.

Questions

Can this tool tell if a photo is AI?

Only when the file says so. Some AI image tools write a label into the file, either as an IPTC digital source type or inside Content Credentials. This tool reads that label. When there is no label, it cannot tell, and neither can anyone else with certainty.

Is my photo uploaded anywhere?

No. The file is read by your browser on your own device. The page sends nothing about it to Rawly or to anyone else, and it keeps working with your connection turned off once the page has loaded.

Why does my photo show no information at all?

Most social apps and messengers strip this data when a photo is posted or sent, to save space and protect privacy. A screenshot also carries none of it. An empty result is normal and says nothing about whether the photo is real.

Does camera data prove a photo is real?

No. Camera make, model and date can be edited or copied from another file with free tools. Treat them as a hint. Content Credentials are harder to fake because they are signed, and you can check the signature in the official Content Credentials verify tool.

How does Rawly keep AI images out?

Rawly does not try to detect AI images. It removes the way in: photos can only be taken with the in-app camera, there are no gallery uploads and no filters, and every shared photo carries a verify code that shows when it was taken in Rawly. It does not show who took it.