AI image detector
Free, no sign-up. your file never leaves your device
How to use
Frequently asked questions
In two layers. First, we look for content signatures (C2PA/Content Credentials) that generators like DALL-E (ChatGPT), Adobe Firefly and Gemini embed in the file — when found, certainty is practically total. Then we examine technical metadata: Stable Diffusion and ComfyUI generation parameters, the IPTC "algorithmic media" marker, software fields and camera data. Everything is read in your browser, without uploading the photo.
When the tool finds a signature or generator metadata, yes — it is direct evidence written inside the file. When nothing is found, nobody in the world can guarantee an answer: WhatsApp, Instagram and screenshots strip metadata, and that is exactly what scammers exploit. In those cases we show "inconclusive" with the weak clues that remain, instead of inventing certainty that does not exist — beware of sites promising "99% accuracy" for any image.
Because the image arrived with no metadata — normal for photos received via WhatsApp, downloaded from social networks or captured as screenshots. If possible, get the original file (ask for it as a "document" on WhatsApp, which preserves the data) and analyze it again: the chance of finding evidence goes way up.
The ones that sign or mark their files: DALL-E and GPT (ChatGPT/OpenAI), Adobe Firefly and Photoshop generative fill, Google Gemini/Imagen, Midjourney, Stable Diffusion (A1111, ComfyUI, InvokeAI, NovelAI), Ideogram, Leonardo, Flux and others. The signature list grows with the market — and photos edited with on-device AI (like Galaxy AI) also carry their own marker.
No. The whole analysis — signatures, metadata and clues — happens inside your browser. The image never leaves your device, and the downloadable report is generated locally too.