AI Image Detector
Check images for AI generation and Content Credentials
Images often carry their own history. AI Inspector Lab reads C2PA Content Credentials, IPTC digital-source-type declarations, EXIF, XMP and PNG text chunks to surface what a file says about how it was made - and runs a detection model when you have one configured.
What the inspector checks
C2PA Content Credentials
Detects embedded manifests in JPEG, PNG and WebP and reads the claim generator, actions and digital source type, such as “trainedAlgorithmicMedia”.
Generator signatures
Recognises generation parameters written by Stable Diffusion WebUI, ComfyUI, InvokeAI and NovelAI, and tool names such as Midjourney, DALL·E and Adobe Firefly.
Camera and software metadata
Reports camera make, model, lens and editing software, with a clear note that EXIF can be copied or forged.
Model likelihood (optional)
When an image detection model is configured, its estimate is shown separately from metadata and provenance evidence.
Who uses it
- Newsrooms verifying user-submitted photos
- Marketplaces checking listing images
- Moderators reviewing profile pictures
- Designers auditing stock assets
AI Image Detector FAQ
Does a C2PA manifest prove an image is AI-generated?
A manifest is a signed claim by the tool that produced or edited the file. We show what it declares and note that we have not verified its signature. Treat it as strong, but checkable, evidence.
Why does my AI image show no indicators?
Re-saving, screenshots and platform uploads usually remove metadata. Without a configured detection model there is nothing left to inspect - and we report that honestly.
Do you decode or alter my image?
No pixels are decoded or modified for metadata inspection. The file is stored privately and deleted after the retention period or when you delete the inspection.
Other detectors
Get the AI Image Detector with Pro
Text, document, image and code inspections are included with Pro and Team - private by default, with Compare and the REST API.