AIorNot

Deepfakes aren't just text anymore: a cloned voice on a scam call, a synthetic candidate in a video interview, a Suno song in a human playlist, an AI photo going viral as breaking news. AI or Not is built for all of it — one detector across text, images, video, voice, and music that tells you not just whether content is AI, but which generator most likely made it.

The text detector scans up to 500,000 characters and highlights the exact AI sentences, color-coded. The image detector reads the pixels to flag fakes and name the model, from Midjourney to Nano Banana. The video detector catches AI-generated and face-swapped clips and can check voice, music, and deepfake signals in one pass. The voice detector separates real speech from clones; the music detector tells human tracks from Suno and Udio songs.

Every verdict comes with a confidence score and the signals behind it, so results are explainable rather than a black box, and the models update as new generators launch. Used by journalists, trust & safety and fraud teams, recruiters, educators, marketplaces, and music pros — plus everyday people checking something before they share it. Free to start, no sign-up.