Detect VALL-E AI voices.
Suspect a clip was made with VALL-E? Run it through the free detector to check whether the voice is AI-generated. VALL-E in short: Microsoft research model that clones a voice from about three seconds of audio using neural codec language modeling.
VALL-E, and why it shows up in suspicious audio
VALL-E is a Microsoft Research text-to-speech model that can reproduce a speaker's voice, and even the acoustic feel of the original recording, from roughly three seconds of sample audio, using a neural codec language-modeling approach. It is a research system rather than a consumer product.
Its approach is influential, and it illustrates how little audio cloning now needs. Most VALL-E clips in the wild are demos or derivatives, and the detector reads the synthesis signature regardless of the exact build.
Where you tend to see it: Research demos, derivative tools, and proof-of-concept impersonation.
VALL-E is a legitimate product; misuse is the problem, not the tool. You can read about it on the official VALL-E site.
How to tell a VALL-E voice
The human ear is unreliable on current VALL-E audio. These are the kinds of artifacts a detector is trained to pick up.
- 01Neural codec artifacts that a model reads even in short clips.
- 02Short-sample cloning that reproduces room tone in a slightly off way.
- 03Prosodic regularity across a passage.
- 04Transitions between phonemes that are cleaner than natural speech.
What the detector tells you
The detector reads the audio, not the speaker. Cloned or stock voice, clean or compressed, it looks for the signs of synthesis.
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. It tells you whether the voice is AI-generated, not which tool made it.
What to do next
A verdict is evidence, not proof. Sign in to keep the result in your detection history and share it with a link. If you are a journalist or investigator, use it alongside your own reporting; if this is a suspected scam, treat the contact as unverified and confirm through a channel you already trust. The FTC's advice on suspected scam calls is a sensible baseline.
Building this into a workflow? The API returns the same verdict as JSON, and a Chrome extension for checking audio in any tab is in development.