Detect a specific AI voice.
Each guide covers one AI voice tool: how it tends to sound, where it shows up, and what to listen for. To check a clip, run it through the free detector, which tells you whether the voice is AI-generated.
Start with the source you suspect
Not sure which tool made a clip? You do not need to know. Drop it into the detector, which checks the audio itself.
Why it helps to know the tools
Different tools tend to show up in different kinds of abuse. A consumer voice-cloning app points to impersonation of a specific person; a cloud text-to-speech engine points to automation or a bot. Knowing what is common helps a fraud analyst, a journalist, or an investigator decide what to check next. The detector itself tells you whether a voice is AI-generated, not which tool made it.
The detector reads the signs of synthesis in the audio rather than the words being said, so cloned and stock voices are both in scope.
What the detector covers
The guides here span the tools people actually encounter: voice-cloning platforms like ElevenLabs, Resemble AI, Cartesia and Descript; studio and read-aloud text-to-speech like Murf, WellSaid and Speechify; the cloud engines from Amazon, Microsoft, Google and OpenAI; and Suno for AI music and vocals. There are also guides to older or discontinued tools, including PlayHT, Replica Studios, Coqui, Bark, Tortoise and Microsoft's VALL-E research model, because clips made with them still circulate. Each guide explains how that tool tends to sound and the tells worth listening for. If your hunch is wrong it does not matter: the detector reads the audio, not your guess. For the underlying method, see how the detector reads a clip; to check something now, open the detector and drop the file in.
The audio, not the speaker
The detector reads the signs of synthesis in the audio, so cloned and stock voices are both in scope. Learn more about deepfake voice detection.
Drop the clip
Upload an MP3, WAV, M4A, OGG, FLAC or AAC file. Three seconds or more of clear speech gives the most reliable result.
The model scores it
The model scores the clip. On longer recordings it scores each segment separately, so you can see which part sounds synthetic.
Read the verdict
You get a verdict on a four-step scale, from Human to AI Generated, plus an AI probability score.